# load styler for styling the R code. Create citations for the thesis report.
# Remaining packages are loaded while coding.
library(styler)
library(report)
citation("report")
citation("styler")
This document can serve as a guideline for performing a social science or marketing study involving video analysis. The features retrieved from the video’s can be used to train and predict classification models.
This document holds the following structure:
Part 1: Preparing the data - Load the processed video CSV file into R - Combining different CSV files into a master data frame - Pre-process the data frame obtained by the loaded and combined data - Data cleaning steps - Feature creation - Creating descriptive statistics
Part 2: EDA - Descriptive statistics tables - Distribution analysis
Part 3: Train and predict - Train and test set - Training models (decision trees and SVM) - Prediction - Evaluation
Part 4: Remain code used in this project
A complete reference list of can be found in the thesis file. References of the packages used to create the code, can be found in the code blocks. They are not printed in the HTML or pdf output. Some ideas have been obtained using help forums like stackoverflow (https://stackoverflow.com/). Usually in the posts you can also find help debugging your code.
The thesis and this RMarkdown code file do not contain any studies with human participants or animals performed by the author. Data used in this study were previously collected. The original owner of the data retains ownership of the data during and after the completion of this thesis.
Before loading the data a new Project directory is created. All data used for this project is stored in this directory.
The database is processed through OpenFace2.0 and consists of \(475\) csv files. These are loaded into R using the dplyr package. The original filename is added to the database using the flnm function. The function appends the filename to each record during the initial reading of the csv files. Next, this function reading multiple csv files add once is used, instead of the read_csv() function.
Attributions and Appreciations: With special thanks to: https://stackoverflow.com/users/5088194/leerssej for providing the code for both functions to append the filenames and load multiple csv files add once in one dataframe.
# load packages
library(tidyverse)
library(data.table)
# *dplyr()*
# `read_csv()`
# functions to load the data
read_plus <- function(flnm) {
read_csv(flnm) %>%
mutate(filename = flnm)
}
map_df_read_csv <- function(path, pattern = "*.csv") {
list.files(path, pattern, full.names = TRUE) %>%
map_df(~ read_plus(.))
}
# create a data frame and relocate the ID column as the first column
df <- map_df_read_csv("Data_Openface/CSV", "*.csv")
df %>% relocate(filename, .before = face_id)
# save the data frame
write.csv(x = df, file = "masterfile", row.names = FALSE)
# citation("tidyverse")
# citation("data.table")
Last step, the dataframe is saved as “masterfile” in the project directory.
Check the filename column using tables of a copy of the original dataframe. First, the dataframe is loaded the dataframe the project drive, and a copy of the file is created. Next, file names are changed into readable files to connect to the gender and age information from separate text files. Open the file of details. Adjust the filename column to be the same as the dataframe. Use inner_join() from the dplyr package to connect the two dataframes. Check the number of observations and variables. Last save the new dataframe as a csv raw database to work with in the next steps of the process, EDA and datacleaning.
From now on the masterfile_connected will be the raw database saved as it contains the full set of elements. The database contains of \(95,830\) observations and \(719\) variables.
# reload the data and create a copy as working file
library(tidyverse)
df <- read.csv("masterfile")
df_copy <- df
df_copy %>% relocate(filename, .before = face_id)
# check the data frame ID
table(df_copy$filename)
# clean filename information to a readable format
df_copy$filename <- gsub(".csv", "", df_copy$filename)
df_copy$filename <- gsub("Data_Openface/CSV/", "", df_copy$filename)
df_copy$filename
table(df_copy$filename)
# open file information from the text file
UvA_subjects <- read.csv("UvA-NEMO_Smile_Database_File_Details.txt",
header = FALSE, skip = 4, sep = "\t"
)
# create column names
colnames(UvA_subjects) <-
c("filename", "subject", "gender", "age", "smile_type")
# clean filename format
UvA_subjects$filename <- gsub(".mp4", "", UvA_subjects$filename)
# join the two datasets by filename
UvA_df <- inner_join(UvA_subjects, df_copy, "filename")
# save the data frame
write.csv(x = UvA_df, file = "masterfile_connected", row.names = FALSE)
For datacleaning, the masterfile_connected is used and will serve as base for further pre-processing. Next step is checking the data for NA’s. The colSums table shows there are no NA values in the database. A visualized NA report is showing no values meaning no NA’s and a full combination field for the dataframe.
# load packages
library(tidyverse)
library(dplyr)
library("VIM")
# load masterfile and create a copy
UvA_original <- read.csv("masterfile_connected")
# check for NA's on different levels using tables and summary data
table(UvA_original$subject, useNA = "always")
table(UvA_original$subject, UvA_original$age)
colSums(is.na(UvA_original))
aggr(UvA_original)
table(UvA_original$age, UvA_original$smile_type)
table(UvA_original$gender, UvA_original$smile_type)
table(UvA_original$smile_type, UvA_original$subject)
summary(UvA_original[, 1:10])
summary(UvA_original[, 11:20])
# AU selection check
summary(UvA_original[, 700:719])
citation("dplyr")
citation("VIM")
Next the confidence and success rate from the original file are checked. All \(95.830\) video frames are processed successfully by OpenFace and the mean() confidence rate is \(98%\).
# Check the confidence and success rate of the loaded video's.
table(UvA_original$confidence)
##
## 0.88 0.93 0.98
## 6 2150 93674
table(UvA_original$success)
##
## 1
## 95830
mean(UvA_original$confidence)
## [1] 0.978872
After loading and checking the data the next step is to create and select the features for the models. There are three different types of features to work with in this research question: parts or symmetry of the face, facial expression and temporal features. In the code below, the dataset is converted to a final dataset including all the features either selected or created. This final dataset is used in the next part of the report to build the models and perform the analysis.In the thesis report an elaboration will be given on the description, selection and calculation of the features.
# load packages
library(readr)
library(scales)
# create a copy from the original dataset to create the final dataset
UvA_final <- UvA_original
# select and create the features for the database using a pipeline
UvA_final <- UvA_final %>%
select(
filename, subject, gender, age, smile_type, frame, timestamp,
starts_with("gaze_angle"), starts_with("pose_"), starts_with("AU"),
x_36, x_37, x_38, x_39, x_42, x_43, x_44, x_45, x_48, x_54, y_36, y_37,
y_38, y_39, y_42, y_43, y_44, y_45, y_48, y_54
) %>%
mutate(AU06_12_c = ifelse(AU06_c == 1 & AU12_c == 1, 1, 0)) %>%
mutate(lip = sqrt((x_48 - x_54)^2 + (y_48 - y_54)^2)) %>%
mutate(eye_x_m_l = (x_36 + x_39) / 2) %>%
mutate(eye_y_m_l = (y_36 + y_39) / 2) %>%
mutate(eye_x_m_r = (x_42 + x_45) / 2) %>%
mutate(eye_y_m_r = (y_42 + y_45) / 2) %>%
mutate(eye_x_u_l = (x_37 + x_38) / 2) %>%
mutate(eye_y_u_l = (y_37 + y_38) / 2) %>%
mutate(eye_x_u_r = (x_43 + x_44) / 2) %>%
mutate(eye_y_u_r = (y_43 + y_44) / 2) %>%
mutate(
eye_l = sqrt((eye_x_m_l - eye_x_u_l)^2 + (eye_y_m_l - eye_y_u_l)^2)
) %>%
mutate(
eye_r = sqrt((eye_x_m_r - eye_x_u_r)^2 + (eye_y_m_r - eye_y_u_r)^2)
) %>%
mutate(eye = (eye_l + eye_r) / 2) %>%
mutate(lip_m_x = (x_48 + x_54) / 2) %>%
mutate(lip_m_y = (y_48 + y_54) / 2) %>%
mutate(amplitude = rescale(sqrt((x_54 - lip_m_x)^2 + (y_54 - lip_m_y)^2))) %>%
mutate(duration = 0.02) %>%
mutate(stage = duration / amplitude) %>%
group_by(filename) %>%
mutate(apex = ifelse(amplitude > mean(amplitude), amplitude, NA)) %>%
mutate(onset_offset = ifelse(amplitude <= mean(amplitude), amplitude, NA)) %>%
mutate(onset = ifelse(onset_offset & frame < mean(frame), amplitude, NA)) %>%
mutate(offset = ifelse(onset_offset & frame >= mean(frame),
amplitude, NA
)) %>%
select(
filename, subject, gender, age, smile_type, frame, timestamp,
starts_with("gaze"), starts_with("pose_R"),
starts_with("AU") & ends_with("_r"),
lip, eye, amplitude, stage, apex, offset, onset
) %>%
ungroup()
# save the final file.
write_csv(UvA_final, "UvA_final")
# Mathematical Notes:
# distance between landmark points 36 and 39 left and 42 and 45 right eyes
# eye middle point of x and y = x_eye_middle (x36 + x39)/2 etc y
# eye upper middle point of x and y = x_eye_upper_middle (x37 + x38)/2
# same for the right eye
# euclidean distance for eye: srqt(x2-x1)^2+(y2-y1)^2
# see calculation lip but than for eye_middle point vs eye_upper_point
# eye gaze only take the angle
# duration ratio = timestamp/duration
# amplitude = lip middle towards left lip corner
# check if sum stats are the same as with OpenFacer - done same result
# all(UvA_face_check$pose_Tz_sd, UvA_sum$pose_Tz_sd)
# citation("readr")
# citation("scales")
For apex, onset and offset, four options for defining the stages have been reviewed. All options show about the same signal as the lip amplitude feature created to measure onset, apex and offset. The options are, the longest subsequence and, start, middle and end point of AU06 and starting point of AU12, based on the categorical variables, and the mean abline() of amplitude. AU12 proves to be always there and starts at different levels for every subject, so this feature could not be used as the categorical variable always displays one. The same goes for the amplitude itself, but because the data show a relative long apex face, the mean is close to the apex value. Therefore a split of stages based on the mean abline() would be a solid solution. The longest subsequence variable could be used as it concerns a single smile but does not show a stable enough sequence to be used (see picture). The AU06 categorical variable starts from zero with every subject, but about 80 participants do not show a offset based on this metric. Therefore the choice is to proceed with the mean abline() split point of the three stages, onset, apex and offset.The plot shows all 4 as an illustration.
Attributions and Appreciations: With special thanks to: Jinjing Xie https://www.r-bloggers.com/2014/09/compute-longest-increasingdecreasing-subsequence-using-rcpp/ for providing the code for the function.
# load packages
library(compiler)
# function for creating longest sub sequence
longest_subseq.R <- cmpfun(function(x) {
P <- integer(length(x))
M <- integer(length(x) + 1)
L <- newL <- 0
for (i in seq_along(x) - 1) {
lo <- 1
hi <- L
while (lo <= hi) {
mid <- (lo + hi) %/% 2
if (x[M[mid + 1] + 1] < x[i + 1]) {
lo <- mid + 1
} else {
hi <- mid - 1
}
}
newL <- lo
P[i + 1] <- M[newL]
if (newL > L) {
M[newL + 1] <- i
L <- newL
} else if (x[i + 1] < x[M[newL + 1] + 1]) {
M[newL + 1] <- i
}
}
k <- M[L + 1]
re <- integer(L)
for (i in L:1) {
re[i] <- k + 1
k <- P[k + 1]
}
re
})
# check the result of one participant number 20.
longest_subseq.R(UvA_final$amplitude[1:193])
## [1] 2 3 5 6 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23
## [20] 24 48 49 50 51 52 53 54 59 60 61 67 68 70 72 78 79 80 83
## [39] 85 90 94 95 100 101 106 108 110 111 113 115 116 119 121
# create graphs to check the four options.
par(mfrow = c(2, 2))
par(mai = c(.8, .8, .2, .2))
# option 1: mean calculation
plot(UvA_final$frame[1:193], UvA_final$amplitude[1:193],
main = "Sequences amplitude",
ylab = "amplitude", xlab = "frame", pch = 16, col = "#0000004D",
cex.main = 0.7
)
abline(mean(UvA_final$amplitude[1:193]), 0, col = "blue", pch = 16)
abline(v = 15, lty = 2)
abline(v = 175, lty = 2)
text(5, 0.7, "onset")
text(100, 0.7, "apex")
text(187, 0.7, "offset")
# option 2: AU12
plot(UvA_final$frame[1:193], UvA_final$AU12_r[1:193],
main = "Sequences AU12_r",
ylab = "amplitude", xlab = "frame", pch = 16, col = "#0000004D",
cex.main = 0.7
)
# option 3: AU06
plot(UvA_final$frame[1:193], UvA_final$AU06_r[1:193],
main = "Sequences AU06_r",
ylab = "amplitude", xlab = "frame", pch = 16, col = "#0000004D",
cex.main = 0.7
)
# option 4: LIS (increasing and decreasing)
plot(UvA_final$amplitude[1:193],
main = "Longest Increasing (blue) and Decreasing (red) Subsequences",
ylab = "amplitude", xlab = "frame", pch = 16, col = "#0000004D",
cex.main = 0.7
)
ind <- longest_subseq.R(UvA_final$amplitude[1:193])
ind
## [1] 2 3 5 6 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23
## [20] 24 48 49 50 51 52 53 54 59 60 61 67 68 70 72 78 79 80 83
## [39] 85 90 94 95 100 101 106 108 110 111 113 115 116 119 121
points(ind, UvA_final$amplitude[1:193][ind], pch = 16, col = "blue")
rind <- longest_subseq.R(-UvA_final$amplitude[1:193])
rind
## [1] 118 121 122 123 125 144 145 146 147 154 155 156 159 164 166 167 169 170 171
## [20] 172 173 174 175 176 177 178 188 190 192 193
points(rind, UvA_final$amplitude[1:193][rind], pch = 16, col = "red")
# citation("compiler")
With the final dataset, summary statistics are created using 2 files. The first file will contain the summary statistics, minimum value min(), maximum value max(), the average value mean() and standard deviation sd(). The min() and max() values will be analyzed from the tables and for the mean() and sd() histograms are made to show the distribution as well as the difference in distribution between the two smile types. This section will be used a set to report these values in the thesis if applicable, e.g., the average time of the video recording or the number of frames. The second file of summary statistics will be used as the input file for models and will contain the mean values only, as described in part 2.
# load packages
library(scales)
library(tidyverse)
library(dplyr)
# create a file to create the summary static files from the the final dataset.
# Using a copy of the final dataset.
UvA_basis <- read.csv("UvA_final")
dim(UvA_basis)
## [1] 95830 36
# file 1 containing all summary statistics for all features.
UvA_sum <- UvA_basis %>%
group_by(filename, subject, gender, age, smile_type) %>%
summarise_all(list(min = min, max = max, mean = mean, sd = sd), na.rm = TRUE)
# save the file
write_csv(UvA_sum, "UvA_sum")
# file 2 create set of features with value mean.
UvA_modelset <- UvA_basis %>%
select(
filename, subject, gender, age, smile_type, gaze_angle_x, gaze_angle_y,
pose_Rx, pose_Ry, pose_Rz, starts_with("AU") & ends_with("_r"),
lip, eye, amplitude, stage, apex, offset, onset
) %>%
group_by(filename, subject, gender, age, smile_type) %>%
summarise_all(list(mean = mean), na.rm = TRUE) %>%
ungroup()
# save the file
write_csv(UvA_modelset, "UvA_modelset")
An overall table of the descriptive statistics is created by the summery() function. For a split based on the classification the psych package is used with the describeBy() function which groups the classifier smile_type and smile_type + gender. Both these files will be analyzed used the output table.
# load packages
library(psych)
# the overall summary of the descriptive statistics
desc_stats <- summary(UvA_sum)
desc_stats
## filename subject gender age
## Length:475 Min. : 20.0 Length:475 Min. : 8.00
## Class :character 1st Qu.:161.5 Class :character 1st Qu.: 9.00
## Mode :character Median :272.0 Mode :character Median :10.00
## Mean :286.7 Mean :10.86
## 3rd Qu.:449.0 3rd Qu.:12.00
## Max. :543.0 Max. :17.00
##
## smile_type frame_min timestamp_min gaze_angle_x_min
## Length:475 Min. :1 Min. :0 Min. :-0.1720
## Class :character 1st Qu.:1 1st Qu.:0 1st Qu.: 0.1040
## Mode :character Median :1 Median :0 Median : 0.1580
## Mean :1 Mean :0 Mean : 0.1478
## 3rd Qu.:1 3rd Qu.:0 3rd Qu.: 0.2000
## Max. :1 Max. :0 Max. : 0.3920
##
## gaze_angle_y_min pose_Rx_min pose_Ry_min pose_Rz_min
## Min. :-0.1260 Min. :-0.3730 Min. :-0.4630 Min. :-0.37500
## 1st Qu.: 0.1485 1st Qu.: 0.0370 1st Qu.:-0.2490 1st Qu.:-0.08500
## Median : 0.2080 Median : 0.1150 Median :-0.2020 Median :-0.03700
## Mean : 0.2041 Mean : 0.1031 Mean :-0.1978 Mean :-0.04409
## 3rd Qu.: 0.2590 3rd Qu.: 0.1760 3rd Qu.:-0.1435 3rd Qu.: 0.00700
## Max. : 0.4840 Max. : 0.3110 Max. : 0.0360 Max. : 0.12900
##
## AU01_r_min AU02_r_min AU04_r_min AU05_r_min AU06_r_min
## Min. :0 Min. :0 Min. :0.0000 Min. :0 Min. :0.00000
## 1st Qu.:0 1st Qu.:0 1st Qu.:0.0000 1st Qu.:0 1st Qu.:0.00000
## Median :0 Median :0 Median :0.0000 Median :0 Median :0.00000
## Mean :0 Mean :0 Mean :0.1395 Mean :0 Mean :0.06373
## 3rd Qu.:0 3rd Qu.:0 3rd Qu.:0.0000 3rd Qu.:0 3rd Qu.:0.00000
## Max. :0 Max. :0 Max. :3.3600 Max. :0 Max. :1.82000
##
## AU07_r_min AU09_r_min AU10_r_min AU12_r_min
## Min. :0.0000 Min. :0 Min. :0.00000 Min. :0.0000
## 1st Qu.:0.0000 1st Qu.:0 1st Qu.:0.00000 1st Qu.:0.0000
## Median :0.0000 Median :0 Median :0.00000 Median :0.0000
## Mean :0.2175 Mean :0 Mean :0.02585 Mean :0.2299
## 3rd Qu.:0.1500 3rd Qu.:0 3rd Qu.:0.00000 3rd Qu.:0.3500
## Max. :3.5900 Max. :0 Max. :1.64000 Max. :2.5600
##
## AU14_r_min AU15_r_min AU17_r_min AU20_r_min AU23_r_min
## Min. :0.000 Min. :0 Min. :0 Min. :0 Min. :0
## 1st Qu.:0.110 1st Qu.:0 1st Qu.:0 1st Qu.:0 1st Qu.:0
## Median :0.510 Median :0 Median :0 Median :0 Median :0
## Mean :0.534 Mean :0 Mean :0 Mean :0 Mean :0
## 3rd Qu.:0.840 3rd Qu.:0 3rd Qu.:0 3rd Qu.:0 3rd Qu.:0
## Max. :1.850 Max. :0 Max. :0 Max. :0 Max. :0
##
## AU25_r_min AU26_r_min AU45_r_min lip_min eye_min
## Min. :0 Min. :0 Min. :0 Min. : 95.03 Min. : 0.3457
## 1st Qu.:0 1st Qu.:0 1st Qu.:0 1st Qu.:136.17 1st Qu.: 2.7174
## Median :0 Median :0 Median :0 Median :148.75 Median : 4.8403
## Mean :0 Mean :0 Mean :0 Mean :148.52 Mean : 5.7299
## 3rd Qu.:0 3rd Qu.:0 3rd Qu.:0 3rd Qu.:159.62 3rd Qu.: 8.7795
## Max. :0 Max. :0 Max. :0 Max. :204.53 Max. :15.4864
##
## amplitude_min stage_min apex_min offset_min
## Min. :0.0000 Min. :0.02000 Min. :0.1344 Min. :0.0006176
## 1st Qu.:0.2745 1st Qu.:0.02731 1st Qu.:0.4488 1st Qu.:0.2973016
## Median :0.3585 Median :0.03148 Median :0.5412 Median :0.3917766
## Mean :0.3569 Mean :0.03383 Mean :0.5388 Mean :0.3886023
## 3rd Qu.:0.4310 3rd Qu.:0.03715 3rd Qu.:0.6427 3rd Qu.:0.4754893
## Max. :0.7306 Max. :0.09461 Max. :0.9014 Max. :0.7306418
##
## onset_min frame_max timestamp_max gaze_angle_x_max
## Min. :0.02589 Min. : 55.0 Min. : 1.080 Min. :0.0170
## 1st Qu.:0.29948 1st Qu.:136.5 1st Qu.: 2.710 1st Qu.:0.1990
## Median :0.38134 Median :176.0 Median : 3.500 Median :0.2480
## Mean :0.37796 Mean :201.7 Mean : 4.015 Mean :0.2441
## 3rd Qu.:0.45648 3rd Qu.:234.0 3rd Qu.: 4.660 3rd Qu.:0.2935
## Max. :0.75256 Max. :705.0 Max. :14.080 Max. :0.4680
##
## gaze_angle_y_max pose_Rx_max pose_Ry_max pose_Rz_max
## Min. :0.0710 Min. :-0.0980 Min. :-0.4020 Min. :-0.15900
## 1st Qu.:0.3325 1st Qu.: 0.1615 1st Qu.:-0.1875 1st Qu.:-0.00750
## Median :0.4040 Median : 0.2180 Median :-0.1320 Median : 0.03700
## Mean :0.4077 Mean : 0.2116 Mean :-0.1313 Mean : 0.04434
## 3rd Qu.:0.4855 3rd Qu.: 0.2805 3rd Qu.:-0.0775 3rd Qu.: 0.08600
## Max. :0.6920 Max. : 0.4460 Max. : 0.1810 Max. : 0.33100
##
## AU01_r_max AU02_r_max AU04_r_max AU05_r_max
## Min. :0.1300 Min. :0.1300 Min. :0.0000 Min. :0.1000
## 1st Qu.:0.4150 1st Qu.:0.3500 1st Qu.:0.0000 1st Qu.:0.3000
## Median :0.6500 Median :0.4700 Median :0.2000 Median :0.4600
## Mean :0.7772 Mean :0.5596 Mean :0.5686 Mean :0.5215
## 3rd Qu.:0.9150 3rd Qu.:0.6300 3rd Qu.:0.8300 3rd Qu.:0.6500
## Max. :5.0000 Max. :3.0900 Max. :4.4900 Max. :2.3400
##
## AU06_r_max AU07_r_max AU09_r_max AU10_r_max
## Min. :0.000 Min. :0.000 Min. :0.0800 Min. :0.000
## 1st Qu.:1.265 1st Qu.:1.165 1st Qu.:0.2100 1st Qu.:0.615
## Median :1.740 Median :1.770 Median :0.2900 Median :1.270
## Mean :1.800 Mean :1.804 Mean :0.3511 Mean :1.210
## 3rd Qu.:2.320 3rd Qu.:2.400 3rd Qu.:0.4400 3rd Qu.:1.740
## Max. :3.740 Max. :5.000 Max. :1.8800 Max. :3.340
##
## AU12_r_max AU14_r_max AU15_r_max AU17_r_max
## Min. :0.790 Min. :0.030 Min. :0.1300 Min. :0.290
## 1st Qu.:2.390 1st Qu.:1.540 1st Qu.:0.2750 1st Qu.:0.835
## Median :2.800 Median :1.850 Median :0.3900 Median :1.200
## Mean :2.782 Mean :1.851 Mean :0.4486 Mean :1.274
## 3rd Qu.:3.200 3rd Qu.:2.170 3rd Qu.:0.5450 3rd Qu.:1.560
## Max. :4.700 Max. :3.430 Max. :2.7600 Max. :3.970
##
## AU20_r_max AU23_r_max AU25_r_max AU26_r_max
## Min. :0.1000 Min. :0.0900 Min. :0.310 Min. :0.290
## 1st Qu.:0.3000 1st Qu.:0.3100 1st Qu.:0.745 1st Qu.:0.780
## Median :0.4200 Median :0.5200 Median :1.140 Median :1.030
## Mean :0.4768 Mean :0.5904 Mean :1.340 Mean :1.143
## 3rd Qu.:0.5900 3rd Qu.:0.7500 3rd Qu.:1.835 3rd Qu.:1.350
## Max. :1.6100 Max. :1.9100 Max. :3.460 Max. :4.730
##
## AU45_r_max lip_max eye_max amplitude_max
## Min. :0.220 Min. :126.7 Min. : 7.248 Min. :0.2114
## 1st Qu.:0.610 1st Qu.:175.7 1st Qu.:11.375 1st Qu.:0.5384
## Median :1.730 Median :190.3 Median :12.624 Median :0.6354
## Mean :1.803 Mean :189.2 Mean :12.747 Mean :0.6282
## 3rd Qu.:2.855 3rd Qu.:204.8 3rd Qu.:14.077 3rd Qu.:0.7323
## Max. :5.000 Max. :244.9 Max. :22.049 Max. :1.0000
##
## stage_max apex_max offset_max onset_max
## Min. :0.02737 Min. :0.2114 Min. :0.1277 Min. :0.1330
## 1st Qu.:0.04640 1st Qu.:0.5384 1st Qu.:0.4402 1st Qu.:0.4401
## Median :0.05579 Median :0.6354 Median :0.5288 Median :0.5301
## Mean : Inf Mean :0.6282 Mean :0.5284 Mean :0.5295
## 3rd Qu.:0.07286 3rd Qu.:0.7323 3rd Qu.:0.6303 3rd Qu.:0.6311
## Max. : Inf Max. :1.0000 Max. :0.8980 Max. :0.8867
##
## frame_mean timestamp_mean gaze_angle_x_mean gaze_angle_y_mean
## Min. : 28.00 Min. :0.540 Min. :-0.03577 Min. :-0.004175
## 1st Qu.: 68.75 1st Qu.:1.355 1st Qu.: 0.15391 1st Qu.: 0.235610
## Median : 88.50 Median :1.750 Median : 0.20332 Median : 0.287323
## Mean :101.37 Mean :2.007 Mean : 0.19704 Mean : 0.282436
## 3rd Qu.:117.50 3rd Qu.:2.330 3rd Qu.: 0.24110 3rd Qu.: 0.329117
## Max. :353.00 Max. :7.040 Max. : 0.42931 Max. : 0.520731
##
## pose_Rx_mean pose_Ry_mean pose_Rz_mean AU01_r_mean
## Min. :-0.1145 Min. :-0.4267 Min. :-0.1801186 Min. :0.03137
## 1st Qu.: 0.1090 1st Qu.:-0.2119 1st Qu.:-0.0458966 1st Qu.:0.06947
## Median : 0.1732 Median :-0.1687 Median :-0.0003729 Median :0.10456
## Mean : 0.1647 Mean :-0.1646 Mean : 0.0001707 Mean :0.13077
## 3rd Qu.: 0.2315 3rd Qu.:-0.1111 3rd Qu.: 0.0429667 3rd Qu.:0.15482
## Max. : 0.3679 Max. : 0.0518 Max. : 0.2370769 Max. :1.06200
##
## AU02_r_mean AU04_r_mean AU05_r_mean AU06_r_mean
## Min. :0.01783 Min. :0.00000 Min. :0.01407 Min. :0.0000
## 1st Qu.:0.03469 1st Qu.:0.00000 1st Qu.:0.02607 1st Qu.:0.6064
## Median :0.04575 Median :0.01016 Median :0.03526 Median :0.9590
## Mean :0.05639 Mean :0.29150 Mean :0.04307 Mean :1.0220
## 3rd Qu.:0.06178 3rd Qu.:0.26155 3rd Qu.:0.05253 3rd Qu.:1.4284
## Max. :0.37496 Max. :3.75359 Max. :0.24129 Max. :2.9500
##
## AU07_r_mean AU09_r_mean AU10_r_mean AU12_r_mean
## Min. :0.0000 Min. :0.01123 Min. :0.0000 Min. :0.3104
## 1st Qu.:0.3844 1st Qu.:0.02544 1st Qu.:0.1689 1st Qu.:1.5263
## Median :0.8927 Median :0.03456 Median :0.5586 Median :1.8963
## Mean :1.0017 Mean :0.04377 Mean :0.6229 Mean :1.8779
## 3rd Qu.:1.4146 3rd Qu.:0.05073 3rd Qu.:0.9638 3rd Qu.:2.2728
## Max. :4.8921 Max. :0.29978 Max. :2.1251 Max. :3.3969
##
## AU14_r_mean AU15_r_mean AU17_r_mean AU20_r_mean
## Min. :0.001011 Min. :0.02597 Min. :0.07526 Min. :0.01914
## 1st Qu.:1.058506 1st Qu.:0.05114 1st Qu.:0.22504 1st Qu.:0.04126
## Median :1.318199 Median :0.06472 Median :0.32262 Median :0.05595
## Mean :1.312488 Mean :0.07568 Mean :0.36455 Mean :0.06591
## 3rd Qu.:1.613258 3rd Qu.:0.08640 3rd Qu.:0.45501 3rd Qu.:0.07602
## Max. :2.805286 Max. :0.36451 Max. :1.67080 Max. :0.31641
##
## AU23_r_mean AU25_r_mean AU26_r_mean AU45_r_mean
## Min. :0.01494 Min. :0.06867 Min. :0.08846 Min. :0.03696
## 1st Qu.:0.04066 1st Qu.:0.21927 1st Qu.:0.19091 1st Qu.:0.08465
## Median :0.06374 Median :0.40466 Median :0.26665 Median :0.14633
## Mean :0.07844 Mean :0.53228 Mean :0.29881 Mean :0.18336
## 3rd Qu.:0.09597 3rd Qu.:0.74531 3rd Qu.:0.36326 3rd Qu.:0.24501
## Max. :0.31894 Max. :1.89141 Max. :1.41270 Max. :0.95777
##
## lip_mean eye_mean amplitude_mean stage_mean
## Min. :115.1 Min. : 4.653 Min. :0.1342 Min. :0.02283
## 1st Qu.:162.0 1st Qu.: 9.114 1st Qu.:0.4468 1st Qu.:0.03222
## Median :175.8 Median :10.381 Median :0.5390 Median :0.03844
## Mean :175.3 Mean :10.466 Mean :0.5358 Mean : Inf
## 3rd Qu.:190.5 3rd Qu.:11.680 3rd Qu.:0.6373 3rd Qu.:0.04669
## Max. :229.6 Max. :17.306 Max. :0.8981 Max. : Inf
##
## apex_mean offset_mean onset_mean frame_sd
## Min. :0.1839 Min. :0.02657 Min. :0.03989 Min. : 16.02
## 1st Qu.:0.5044 1st Qu.:0.35351 1st Qu.:0.34981 1st Qu.: 39.55
## Median :0.6008 Median :0.44376 Median :0.43873 Median : 50.95
## Mean :0.5959 Mean :0.44482 Mean :0.43614 Mean : 58.38
## 3rd Qu.:0.6972 3rd Qu.:0.54219 3rd Qu.:0.52342 3rd Qu.: 67.69
## Max. :0.9582 Max. :0.77740 Max. :0.78238 Max. :203.66
##
## timestamp_sd gaze_angle_x_sd gaze_angle_y_sd pose_Rx_sd
## Min. :0.3204 Min. :0.004268 Min. :0.006175 Min. :0.005344
## 1st Qu.:0.7910 1st Qu.:0.012138 1st Qu.:0.026885 1st Qu.:0.015464
## Median :1.0190 Median :0.016913 Median :0.037900 Median :0.023031
## Mean :1.1677 Mean :0.020123 Mean :0.040922 Mean :0.027226
## 3rd Qu.:1.3539 3rd Qu.:0.023455 3rd Qu.:0.051110 3rd Qu.:0.032555
## Max. :4.0732 Max. :0.101646 Max. :0.122763 Max. :0.156937
##
## pose_Ry_sd pose_Rz_sd AU01_r_sd AU02_r_sd
## Min. :0.003045 Min. :0.002531 Min. :0.03744 Min. :0.03076
## 1st Qu.:0.008281 1st Qu.:0.009640 1st Qu.:0.10330 1st Qu.:0.06795
## Median :0.012707 Median :0.016293 Median :0.16113 Median :0.09569
## Mean :0.016447 Mean :0.023986 Mean :0.20228 Mean :0.11850
## 3rd Qu.:0.020009 3rd Qu.:0.028439 3rd Qu.:0.24067 3rd Qu.:0.12989
## Max. :0.094527 Max. :0.165375 Max. :1.77870 Max. :0.79933
##
## AU04_r_sd AU05_r_sd AU06_r_sd AU07_r_sd
## Min. :0.00000 Min. :0.02550 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.00000 1st Qu.:0.05897 1st Qu.:0.4205 1st Qu.:0.2867
## Median :0.03379 Median :0.08229 Median :0.5532 Median :0.4026
## Mean :0.10454 Mean :0.10411 Mean :0.5699 Mean :0.4193
## 3rd Qu.:0.17489 3rd Qu.:0.12720 3rd Qu.:0.7176 3rd Qu.:0.5393
## Max. :0.67294 Max. :0.60876 Max. :1.3138 Max. :1.2569
##
## AU09_r_sd AU10_r_sd AU12_r_sd AU14_r_sd
## Min. :0.01757 Min. :0.0000 Min. :0.2404 Min. :0.00543
## 1st Qu.:0.04715 1st Qu.:0.1979 1st Qu.:0.6376 1st Qu.:0.22035
## Median :0.06596 Median :0.3946 Median :0.7956 Median :0.29707
## Mean :0.08367 Mean :0.3789 Mean :0.8276 Mean :0.31989
## 3rd Qu.:0.10112 3rd Qu.:0.5431 3rd Qu.:0.9779 3rd Qu.:0.40174
## Max. :0.56374 Max. :1.1471 Max. :2.0527 Max. :0.82670
##
## AU15_r_sd AU17_r_sd AU20_r_sd AU23_r_sd
## Min. :0.03162 Min. :0.06655 Min. :0.02474 Min. :0.02427
## 1st Qu.:0.07073 1st Qu.:0.21955 1st Qu.:0.06895 1st Qu.:0.07299
## Median :0.09319 Median :0.33761 Median :0.09610 Median :0.12103
## Mean :0.11180 Mean :0.37142 Mean :0.11509 Mean :0.14534
## 3rd Qu.:0.13323 3rd Qu.:0.45936 3rd Qu.:0.14000 3rd Qu.:0.19001
## Max. :0.64167 Max. :1.49028 Max. :0.52165 Max. :0.54104
##
## AU25_r_sd AU26_r_sd AU45_r_sd lip_sd
## Min. :0.06876 Min. :0.06916 Min. :0.04699 Min. : 2.839
## 1st Qu.:0.20905 1st Qu.:0.21161 1st Qu.:0.14292 1st Qu.: 9.507
## Median :0.37274 Median :0.29238 Median :0.31528 Median :11.928
## Mean :0.46673 Mean :0.32550 Mean :0.37306 Mean :12.429
## 3rd Qu.:0.67814 3rd Qu.:0.39434 3rd Qu.:0.55286 3rd Qu.:15.330
## Max. :1.43521 Max. :1.46384 Max. :1.50215 Max. :31.297
##
## eye_sd amplitude_sd stage_sd apex_sd
## Min. :0.1934 Min. :0.01894 Min. :0.001363 Min. :0.005714
## 1st Qu.:0.8221 1st Qu.:0.06344 1st Qu.:0.004642 1st Qu.:0.015893
## Median :1.3406 Median :0.07959 Median :0.006758 Median :0.021934
## Mean :1.4573 Mean :0.08293 Mean :0.009031 Mean :0.022950
## 3rd Qu.:1.9303 3rd Qu.:0.10229 3rd Qu.:0.010379 3rd Qu.:0.028079
## Max. :5.1009 Max. :0.20884 Max. :0.164760 Max. :0.074245
## NA's :1
## offset_sd onset_sd
## Min. :0.001877 Min. :0.002385
## 1st Qu.:0.026013 1st Qu.:0.031080
## Median :0.038472 Median :0.045371
## Mean :0.041893 Mean :0.048238
## 3rd Qu.:0.055286 3rd Qu.:0.062726
## Max. :0.139375 Max. :0.153071
##
# descriptive statistics per classifier
desc_stats_class <- describeBy(UvA_sum, UvA_sum$smile_type)
desc_stats_class
##
## Descriptive statistics by group
## group: deliberate
## vars n mean sd median trimmed mad min max
## filename* 1 240 120.50 69.43 120.50 120.50 88.96 1.00 240.00
## subject 2 240 290.22 154.82 277.00 289.55 214.98 20.00 543.00
## gender* 3 240 1.52 0.50 2.00 1.52 0.00 1.00 2.00
## age 4 240 10.96 2.36 10.50 10.70 2.22 8.00 17.00
## smile_type* 5 240 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## frame_min 6 240 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## timestamp_min 7 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## gaze_angle_x_min 8 240 0.14 0.08 0.15 0.15 0.07 -0.17 0.37
## gaze_angle_y_min 9 240 0.21 0.09 0.22 0.21 0.08 -0.10 0.48
## pose_Rx_min 10 240 0.11 0.10 0.13 0.12 0.11 -0.33 0.29
## pose_Ry_min 11 240 -0.19 0.08 -0.20 -0.19 0.08 -0.44 0.03
## pose_Rz_min 12 240 -0.04 0.07 -0.03 -0.04 0.06 -0.26 0.09
## AU01_r_min 13 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU02_r_min 14 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU04_r_min 15 240 0.14 0.47 0.00 0.01 0.00 0.00 2.99
## AU05_r_min 16 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU06_r_min 17 240 0.05 0.19 0.00 0.00 0.00 0.00 1.35
## AU07_r_min 18 240 0.19 0.49 0.00 0.06 0.00 0.00 3.46
## AU09_r_min 19 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU10_r_min 20 240 0.02 0.13 0.00 0.00 0.00 0.00 1.64
## AU12_r_min 21 240 0.16 0.32 0.00 0.09 0.00 0.00 2.56
## AU14_r_min 22 240 0.51 0.44 0.47 0.47 0.53 0.00 1.82
## AU15_r_min 23 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU17_r_min 24 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU20_r_min 25 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU23_r_min 26 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU25_r_min 27 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU26_r_min 28 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU45_r_min 29 240 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## lip_min 30 240 147.95 18.35 147.61 147.84 17.94 95.03 204.53
## eye_min 31 240 5.79 3.64 5.06 5.53 4.20 0.54 15.49
## amplitude_min 32 240 0.35 0.12 0.35 0.35 0.12 0.00 0.73
## stage_min 33 240 0.03 0.01 0.03 0.03 0.01 0.02 0.09
## apex_min 34 240 0.55 0.13 0.54 0.55 0.14 0.13 0.84
## offset_min 35 240 0.38 0.13 0.38 0.38 0.13 0.00 0.73
## onset_min 36 240 0.38 0.12 0.38 0.38 0.11 0.03 0.75
## frame_max 37 240 159.69 51.34 151.00 152.72 37.81 84.00 438.00
## timestamp_max 38 240 3.17 1.03 3.00 3.03 0.76 1.66 8.74
## gaze_angle_x_max 39 240 0.24 0.07 0.24 0.24 0.07 0.02 0.46
## gaze_angle_y_max 40 240 0.41 0.11 0.40 0.41 0.11 0.10 0.69
## pose_Rx_max 41 240 0.21 0.10 0.21 0.21 0.09 -0.10 0.42
## pose_Ry_max 42 240 -0.13 0.08 -0.13 -0.13 0.08 -0.38 0.18
## pose_Rz_max 43 240 0.04 0.08 0.03 0.03 0.06 -0.16 0.33
## AU01_r_max 44 240 0.82 0.60 0.72 0.73 0.42 0.13 5.00
## AU02_r_max 45 240 0.53 0.33 0.47 0.47 0.21 0.13 3.09
## AU04_r_max 46 240 0.50 0.76 0.12 0.34 0.18 0.00 4.10
## AU05_r_max 47 240 0.53 0.35 0.44 0.47 0.25 0.10 2.34
## AU06_r_max 48 240 1.71 0.76 1.68 1.71 0.74 0.00 3.53
## AU07_r_max 49 240 1.67 0.93 1.62 1.64 0.88 0.00 5.00
## AU09_r_max 50 240 0.32 0.17 0.27 0.29 0.13 0.08 0.98
## AU10_r_max 51 240 1.05 0.74 1.07 1.02 0.87 0.00 2.86
## AU12_r_max 52 240 2.77 0.67 2.74 2.78 0.66 0.79 4.70
## AU14_r_max 53 240 1.86 0.43 1.85 1.86 0.47 0.83 2.89
## AU15_r_max 54 240 0.42 0.23 0.36 0.39 0.16 0.13 1.27
## AU17_r_max 55 240 1.25 0.64 1.19 1.17 0.55 0.29 3.97
## AU20_r_max 56 240 0.45 0.24 0.39 0.42 0.19 0.10 1.46
## AU23_r_max 57 240 0.56 0.35 0.48 0.51 0.29 0.09 1.88
## AU25_r_max 58 240 1.40 0.80 1.22 1.34 0.85 0.31 3.38
## AU26_r_max 59 240 1.10 0.48 1.02 1.05 0.42 0.31 2.93
## AU45_r_max 60 240 1.91 1.31 1.91 1.84 1.85 0.22 5.00
## lip_max 61 240 190.26 21.01 191.42 190.83 22.46 126.71 239.26
## eye_max 62 240 12.92 2.15 12.78 12.88 1.78 7.67 19.03
## amplitude_max 63 240 0.64 0.14 0.64 0.64 0.15 0.21 0.96
## stage_max 64 240 Inf NaN 0.06 0.06 0.02 0.03 Inf
## apex_max 65 240 0.64 0.14 0.64 0.64 0.15 0.21 0.96
## offset_max 66 240 0.53 0.13 0.53 0.53 0.14 0.13 0.84
## onset_max 67 240 0.53 0.13 0.53 0.54 0.14 0.13 0.84
## frame_mean 68 240 80.34 25.67 76.00 76.86 18.90 42.50 219.50
## timestamp_mean 69 240 1.59 0.51 1.50 1.52 0.38 0.83 4.37
## gaze_angle_x_mean 70 240 0.19 0.07 0.20 0.19 0.06 -0.04 0.41
## gaze_angle_y_mean 71 240 0.29 0.08 0.29 0.29 0.07 0.05 0.52
## pose_Rx_mean 72 240 0.17 0.10 0.17 0.17 0.10 -0.11 0.37
## pose_Ry_mean 73 240 -0.16 0.08 -0.17 -0.16 0.07 -0.41 0.05
## pose_Rz_mean 74 240 0.00 0.06 0.00 0.00 0.06 -0.18 0.24
## AU01_r_mean 75 240 0.14 0.13 0.12 0.12 0.07 0.03 1.06
## AU02_r_mean 76 240 0.06 0.04 0.05 0.05 0.02 0.02 0.37
## AU04_r_mean 77 240 0.27 0.59 0.01 0.12 0.01 0.00 3.50
## AU05_r_mean 78 240 0.05 0.03 0.04 0.04 0.02 0.01 0.24
## AU06_r_mean 79 240 0.97 0.58 0.91 0.94 0.53 0.00 2.82
## AU07_r_mean 80 240 0.93 0.78 0.80 0.83 0.67 0.00 4.89
## AU09_r_mean 81 240 0.04 0.02 0.03 0.04 0.02 0.01 0.15
## AU10_r_mean 82 240 0.53 0.48 0.48 0.47 0.53 0.00 2.06
## AU12_r_mean 83 240 1.84 0.57 1.86 1.85 0.56 0.31 3.34
## AU14_r_mean 84 240 1.33 0.41 1.32 1.34 0.42 0.31 2.20
## AU15_r_mean 85 240 0.07 0.04 0.06 0.07 0.03 0.03 0.36
## AU17_r_mean 86 240 0.37 0.20 0.34 0.35 0.17 0.08 1.67
## AU20_r_mean 87 240 0.06 0.04 0.05 0.06 0.03 0.02 0.27
## AU23_r_mean 88 240 0.08 0.06 0.06 0.07 0.04 0.02 0.29
## AU25_r_mean 89 240 0.61 0.44 0.49 0.56 0.43 0.07 1.89
## AU26_r_mean 90 240 0.29 0.14 0.26 0.27 0.11 0.09 1.13
## AU45_r_mean 91 240 0.20 0.14 0.17 0.18 0.13 0.04 0.96
## lip_mean 92 240 176.25 19.78 175.91 176.85 20.89 115.14 220.98
## eye_mean 93 240 10.64 2.04 10.53 10.60 1.56 4.65 16.57
## amplitude_mean 94 240 0.54 0.13 0.54 0.55 0.14 0.13 0.84
## stage_mean 95 240 Inf NaN 0.04 0.04 0.01 0.02 Inf
## apex_mean 96 240 0.61 0.14 0.61 0.61 0.15 0.19 0.89
## offset_mean 97 240 0.43 0.13 0.43 0.43 0.14 0.03 0.78
## onset_mean 98 240 0.43 0.12 0.43 0.43 0.13 0.04 0.78
## frame_sd 99 240 46.24 14.82 43.73 44.23 10.91 24.39 126.58
## timestamp_sd 100 240 0.92 0.30 0.87 0.88 0.22 0.49 2.53
## gaze_angle_x_sd 101 240 0.02 0.01 0.02 0.02 0.01 0.00 0.10
## gaze_angle_y_sd 102 240 0.04 0.02 0.04 0.04 0.02 0.01 0.12
## pose_Rx_sd 103 240 0.02 0.01 0.02 0.02 0.01 0.01 0.08
## pose_Ry_sd 104 240 0.02 0.01 0.01 0.01 0.01 0.00 0.07
## pose_Rz_sd 105 240 0.02 0.02 0.01 0.02 0.01 0.00 0.14
## AU01_r_sd 106 240 0.22 0.20 0.18 0.19 0.11 0.04 1.78
## AU02_r_sd 107 240 0.12 0.09 0.10 0.10 0.04 0.03 0.80
## AU04_r_sd 108 240 0.09 0.12 0.02 0.07 0.03 0.00 0.59
## AU05_r_sd 109 240 0.11 0.08 0.08 0.10 0.05 0.03 0.61
## AU06_r_sd 110 240 0.59 0.27 0.57 0.59 0.24 0.00 1.31
## AU07_r_sd 111 240 0.42 0.23 0.40 0.41 0.20 0.00 1.15
## AU09_r_sd 112 240 0.08 0.04 0.06 0.07 0.04 0.02 0.25
## AU10_r_sd 113 240 0.36 0.27 0.34 0.35 0.33 0.00 1.12
## AU12_r_sd 114 240 0.92 0.28 0.89 0.90 0.29 0.24 2.05
## AU14_r_sd 115 240 0.34 0.13 0.32 0.33 0.13 0.12 0.82
## AU15_r_sd 116 240 0.11 0.06 0.09 0.10 0.04 0.03 0.46
## AU17_r_sd 117 240 0.39 0.23 0.35 0.36 0.20 0.07 1.49
## AU20_r_sd 118 240 0.11 0.07 0.09 0.10 0.05 0.02 0.42
## AU23_r_sd 119 240 0.14 0.10 0.12 0.13 0.08 0.02 0.51
## AU25_r_sd 120 240 0.51 0.35 0.41 0.47 0.35 0.07 1.44
## AU26_r_sd 121 240 0.32 0.16 0.29 0.30 0.14 0.09 1.08
## AU45_r_sd 122 240 0.41 0.30 0.36 0.39 0.33 0.05 1.50
## lip_sd 123 240 13.83 4.38 13.57 13.70 4.12 2.84 31.30
## eye_sd 124 240 1.53 0.86 1.40 1.46 0.91 0.19 5.10
## amplitude_sd 125 240 0.09 0.03 0.09 0.09 0.03 0.02 0.21
## stage_sd 126 239 0.01 0.01 0.01 0.01 0.00 0.00 0.16
## apex_sd 127 240 0.02 0.01 0.02 0.02 0.01 0.01 0.07
## offset_sd 128 240 0.04 0.02 0.04 0.04 0.02 0.00 0.13
## onset_sd 129 240 0.05 0.02 0.05 0.05 0.02 0.00 0.15
## range skew kurtosis se
## filename* 239.00 0.00 -1.22 4.48
## subject 523.00 0.09 -1.40 9.99
## gender* 1.00 -0.07 -2.00 0.03
## age 9.00 0.85 0.00 0.15
## smile_type* 0.00 NaN NaN 0.00
## frame_min 0.00 NaN NaN 0.00
## timestamp_min 0.00 NaN NaN 0.00
## gaze_angle_x_min 0.54 -0.84 1.66 0.01
## gaze_angle_y_min 0.58 -0.27 0.51 0.01
## pose_Rx_min 0.62 -0.70 0.64 0.01
## pose_Ry_min 0.47 0.04 0.28 0.01
## pose_Rz_min 0.36 -0.59 0.07 0.00
## AU01_r_min 0.00 NaN NaN 0.00
## AU02_r_min 0.00 NaN NaN 0.00
## AU04_r_min 2.99 4.16 17.86 0.03
## AU05_r_min 0.00 NaN NaN 0.00
## AU06_r_min 1.35 4.68 23.81 0.01
## AU07_r_min 3.46 3.34 13.33 0.03
## AU09_r_min 0.00 NaN NaN 0.00
## AU10_r_min 1.64 9.84 105.19 0.01
## AU12_r_min 2.56 3.06 13.73 0.02
## AU14_r_min 1.82 0.53 -0.54 0.03
## AU15_r_min 0.00 NaN NaN 0.00
## AU17_r_min 0.00 NaN NaN 0.00
## AU20_r_min 0.00 NaN NaN 0.00
## AU23_r_min 0.00 NaN NaN 0.00
## AU25_r_min 0.00 NaN NaN 0.00
## AU26_r_min 0.00 NaN NaN 0.00
## AU45_r_min 0.00 NaN NaN 0.00
## lip_min 109.50 0.07 -0.02 1.18
## eye_min 14.95 0.53 -0.75 0.24
## amplitude_min 0.73 0.07 -0.02 0.01
## stage_min 0.07 2.30 9.11 0.00
## apex_min 0.71 -0.28 -0.29 0.01
## offset_min 0.73 0.02 -0.20 0.01
## onset_min 0.73 0.00 0.12 0.01
## frame_max 354.00 1.90 5.55 3.31
## timestamp_max 7.08 1.90 5.55 0.07
## gaze_angle_x_max 0.44 -0.17 0.35 0.00
## gaze_angle_y_max 0.59 0.20 -0.17 0.01
## pose_Rx_max 0.52 -0.42 -0.09 0.01
## pose_Ry_max 0.56 0.27 0.88 0.01
## pose_Rz_max 0.49 0.89 1.87 0.01
## AU01_r_max 4.87 2.96 13.58 0.04
## AU02_r_max 2.96 3.28 17.31 0.02
## AU04_r_max 4.10 2.17 5.31 0.05
## AU05_r_max 2.24 2.08 6.26 0.02
## AU06_r_max 3.53 0.06 -0.34 0.05
## AU07_r_max 5.00 0.47 0.61 0.06
## AU09_r_max 0.90 1.26 1.38 0.01
## AU10_r_max 2.86 0.13 -0.97 0.05
## AU12_r_max 3.91 -0.09 0.10 0.04
## AU14_r_max 2.06 -0.09 -0.56 0.03
## AU15_r_max 1.14 1.58 2.41 0.01
## AU17_r_max 3.68 1.36 2.79 0.04
## AU20_r_max 1.36 1.32 1.90 0.02
## AU23_r_max 1.79 1.22 1.20 0.02
## AU25_r_max 3.07 0.59 -0.69 0.05
## AU26_r_max 2.62 1.04 1.10 0.03
## AU45_r_max 4.78 0.27 -1.22 0.08
## lip_max 112.55 -0.23 -0.18 1.36
## eye_max 11.36 0.21 0.15 0.14
## amplitude_max 0.75 -0.23 -0.18 0.01
## stage_max Inf NaN NaN NaN
## apex_max 0.75 -0.23 -0.18 0.01
## offset_max 0.71 -0.21 -0.30 0.01
## onset_max 0.70 -0.27 -0.29 0.01
## frame_mean 177.00 1.90 5.55 1.66
## timestamp_mean 3.54 1.90 5.55 0.03
## gaze_angle_x_mean 0.45 -0.30 0.73 0.00
## gaze_angle_y_mean 0.47 -0.11 0.36 0.01
## pose_Rx_mean 0.48 -0.47 -0.18 0.01
## pose_Ry_mean 0.45 0.09 0.33 0.00
## pose_Rz_mean 0.42 0.04 0.55 0.00
## AU01_r_mean 1.03 4.32 24.41 0.01
## AU02_r_mean 0.36 4.22 23.97 0.00
## AU04_r_mean 3.50 3.22 11.30 0.04
## AU05_r_mean 0.23 2.64 10.95 0.00
## AU06_r_mean 2.82 0.55 -0.08 0.04
## AU07_r_mean 4.89 1.43 3.47 0.05
## AU09_r_mean 0.14 1.49 2.66 0.00
## AU10_r_mean 2.06 0.82 0.16 0.03
## AU12_r_mean 3.03 -0.04 -0.19 0.04
## AU14_r_mean 1.89 -0.19 -0.62 0.03
## AU15_r_mean 0.34 2.78 13.00 0.00
## AU17_r_mean 1.60 1.90 7.24 0.01
## AU20_r_mean 0.26 1.98 4.95 0.00
## AU23_r_mean 0.27 1.51 2.08 0.00
## AU25_r_mean 1.82 0.84 -0.17 0.03
## AU26_r_mean 1.04 1.86 6.04 0.01
## AU45_r_mean 0.92 1.58 4.24 0.01
## lip_mean 105.84 -0.26 -0.30 1.28
## eye_mean 11.92 0.25 0.52 0.13
## amplitude_mean 0.71 -0.26 -0.30 0.01
## stage_mean Inf NaN NaN NaN
## apex_mean 0.70 -0.24 -0.33 0.01
## offset_mean 0.75 -0.07 -0.21 0.01
## onset_mean 0.74 -0.18 0.08 0.01
## frame_sd 102.19 1.90 5.55 0.96
## timestamp_sd 2.04 1.90 5.55 0.02
## gaze_angle_x_sd 0.10 2.65 9.71 0.00
## gaze_angle_y_sd 0.11 1.07 1.99 0.00
## pose_Rx_sd 0.08 1.52 2.72 0.00
## pose_Ry_sd 0.07 2.36 6.50 0.00
## pose_Rz_sd 0.14 2.30 5.81 0.00
## AU01_r_sd 1.74 4.10 23.08 0.01
## AU02_r_sd 0.77 3.81 20.22 0.01
## AU04_r_sd 0.59 1.49 1.80 0.01
## AU05_r_sd 0.58 2.40 8.93 0.01
## AU06_r_sd 1.31 0.04 -0.30 0.02
## AU07_r_sd 1.15 0.28 -0.07 0.02
## AU09_r_sd 0.23 1.28 1.28 0.00
## AU10_r_sd 1.12 0.30 -0.82 0.02
## AU12_r_sd 1.81 0.61 0.97 0.02
## AU14_r_sd 0.70 0.62 0.14 0.01
## AU15_r_sd 0.43 2.07 5.60 0.00
## AU17_r_sd 1.42 1.56 3.86 0.01
## AU20_r_sd 0.39 1.72 3.08 0.00
## AU23_r_sd 0.49 1.32 1.34 0.01
## AU25_r_sd 1.37 0.74 -0.45 0.02
## AU26_r_sd 1.00 1.32 2.53 0.01
## AU45_r_sd 1.46 0.79 0.27 0.02
## lip_sd 28.46 0.53 1.09 0.28
## eye_sd 4.91 0.87 1.10 0.06
## amplitude_sd 0.19 0.53 1.09 0.00
## stage_sd 0.16 9.06 109.53 0.00
## apex_sd 0.07 1.45 5.25 0.00
## offset_sd 0.13 0.80 0.97 0.00
## onset_sd 0.15 0.71 0.90 0.00
## ------------------------------------------------------------
## group: spontaneous
## vars n mean sd median trimmed mad min max
## filename* 1 235 118.00 67.98 118.00 118.00 87.47 1.00 235.00
## subject 2 235 283.17 148.56 258.00 280.91 195.70 20.00 534.00
## gender* 3 235 1.52 0.50 2.00 1.52 0.00 1.00 2.00
## age 4 235 10.76 2.30 10.00 10.49 1.48 8.00 17.00
## smile_type* 5 235 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## frame_min 6 235 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## timestamp_min 7 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## gaze_angle_x_min 8 235 0.15 0.08 0.17 0.16 0.07 -0.13 0.39
## gaze_angle_y_min 9 235 0.20 0.09 0.20 0.20 0.08 -0.13 0.48
## pose_Rx_min 10 235 0.09 0.11 0.11 0.10 0.10 -0.37 0.31
## pose_Ry_min 11 235 -0.20 0.09 -0.20 -0.20 0.08 -0.46 0.04
## pose_Rz_min 12 235 -0.05 0.08 -0.04 -0.04 0.08 -0.38 0.13
## AU01_r_min 13 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU02_r_min 14 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU04_r_min 15 235 0.14 0.45 0.00 0.02 0.00 0.00 3.36
## AU05_r_min 16 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU06_r_min 17 235 0.08 0.23 0.00 0.02 0.00 0.00 1.82
## AU07_r_min 18 235 0.24 0.52 0.00 0.11 0.00 0.00 3.59
## AU09_r_min 19 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU10_r_min 20 235 0.03 0.13 0.00 0.00 0.00 0.00 0.86
## AU12_r_min 21 235 0.30 0.43 0.05 0.21 0.07 0.00 1.95
## AU14_r_min 22 235 0.56 0.44 0.53 0.52 0.53 0.00 1.85
## AU15_r_min 23 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU17_r_min 24 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU20_r_min 25 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU23_r_min 26 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU25_r_min 27 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU26_r_min 28 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU45_r_min 29 235 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## lip_min 30 235 149.10 17.74 149.13 148.93 17.14 101.23 189.97
## eye_min 31 235 5.67 3.47 4.70 5.43 3.65 0.35 15.45
## amplitude_min 32 235 0.36 0.12 0.36 0.36 0.11 0.04 0.63
## stage_min 33 235 0.03 0.01 0.03 0.03 0.01 0.02 0.09
## apex_min 34 235 0.53 0.13 0.54 0.53 0.14 0.16 0.90
## offset_min 35 235 0.40 0.12 0.40 0.40 0.13 0.09 0.65
## onset_min 36 235 0.38 0.12 0.38 0.38 0.12 0.04 0.74
## frame_max 37 235 244.70 105.60 224.00 234.26 93.40 55.00 705.00
## timestamp_max 38 235 4.87 2.11 4.46 4.67 1.87 1.08 14.08
## gaze_angle_x_max 39 235 0.25 0.07 0.25 0.25 0.07 0.02 0.47
## gaze_angle_y_max 40 235 0.40 0.11 0.41 0.41 0.12 0.07 0.62
## pose_Rx_max 41 235 0.22 0.09 0.22 0.22 0.08 -0.06 0.45
## pose_Ry_max 42 235 -0.13 0.09 -0.13 -0.13 0.09 -0.40 0.11
## pose_Rz_max 43 235 0.05 0.08 0.05 0.04 0.07 -0.11 0.30
## AU01_r_max 44 235 0.73 0.53 0.59 0.64 0.34 0.15 4.52
## AU02_r_max 45 235 0.59 0.40 0.46 0.51 0.22 0.16 2.24
## AU04_r_max 46 235 0.64 0.87 0.29 0.46 0.43 0.00 4.49
## AU05_r_max 47 235 0.51 0.27 0.46 0.48 0.24 0.12 2.01
## AU06_r_max 48 235 1.89 0.77 1.85 1.88 0.74 0.00 3.74
## AU07_r_max 49 235 1.94 0.99 1.89 1.90 0.92 0.00 5.00
## AU09_r_max 50 235 0.39 0.25 0.32 0.35 0.16 0.08 1.88
## AU10_r_max 51 235 1.37 0.74 1.40 1.37 0.76 0.00 3.34
## AU12_r_max 52 235 2.79 0.59 2.84 2.82 0.58 1.19 4.20
## AU14_r_max 53 235 1.85 0.45 1.86 1.85 0.46 0.03 3.43
## AU15_r_max 54 235 0.47 0.27 0.41 0.44 0.18 0.14 2.76
## AU17_r_max 55 235 1.30 0.60 1.21 1.24 0.52 0.30 3.96
## AU20_r_max 56 235 0.50 0.24 0.44 0.47 0.22 0.12 1.61
## AU23_r_max 57 235 0.63 0.38 0.57 0.58 0.34 0.10 1.91
## AU25_r_max 58 235 1.28 0.70 1.09 1.20 0.67 0.32 3.46
## AU26_r_max 59 235 1.18 0.60 1.07 1.10 0.42 0.29 4.73
## AU45_r_max 60 235 1.69 1.09 1.54 1.63 1.45 0.22 4.27
## lip_max 61 235 188.09 20.82 189.16 188.57 22.50 127.31 244.90
## eye_max 62 235 12.57 2.32 12.55 12.53 2.15 7.25 22.05
## amplitude_max 63 235 0.62 0.14 0.63 0.62 0.15 0.22 1.00
## stage_max 64 235 0.06 0.04 0.06 0.06 0.02 0.03 0.48
## apex_max 65 235 0.62 0.14 0.63 0.62 0.15 0.22 1.00
## offset_max 66 235 0.53 0.13 0.53 0.53 0.15 0.15 0.90
## onset_max 67 235 0.53 0.13 0.53 0.53 0.14 0.16 0.89
## frame_mean 68 235 122.85 52.80 112.50 117.63 46.70 28.00 353.00
## timestamp_mean 69 235 2.44 1.06 2.23 2.33 0.93 0.54 7.04
## gaze_angle_x_mean 70 235 0.20 0.07 0.21 0.20 0.06 -0.01 0.43
## gaze_angle_y_mean 71 235 0.28 0.08 0.28 0.28 0.07 0.00 0.50
## pose_Rx_mean 72 235 0.16 0.09 0.17 0.17 0.08 -0.11 0.37
## pose_Ry_mean 73 235 -0.17 0.08 -0.17 -0.17 0.07 -0.43 0.05
## pose_Rz_mean 74 235 0.00 0.07 0.00 0.00 0.07 -0.18 0.19
## AU01_r_mean 75 235 0.12 0.09 0.10 0.10 0.05 0.03 0.95
## AU02_r_mean 76 235 0.06 0.04 0.04 0.05 0.02 0.02 0.29
## AU04_r_mean 77 235 0.31 0.63 0.02 0.16 0.02 0.00 3.75
## AU05_r_mean 78 235 0.04 0.02 0.03 0.04 0.01 0.01 0.12
## AU06_r_mean 79 235 1.08 0.62 0.99 1.04 0.65 0.00 2.95
## AU07_r_mean 80 235 1.08 0.83 0.96 1.00 0.85 0.00 4.76
## AU09_r_mean 81 235 0.05 0.04 0.04 0.04 0.02 0.01 0.30
## AU10_r_mean 82 235 0.72 0.52 0.67 0.70 0.61 0.00 2.13
## AU12_r_mean 83 235 1.91 0.58 1.92 1.92 0.54 0.40 3.40
## AU14_r_mean 84 235 1.30 0.42 1.31 1.31 0.42 0.00 2.81
## AU15_r_mean 85 235 0.08 0.04 0.07 0.07 0.03 0.03 0.30
## AU17_r_mean 86 235 0.36 0.19 0.31 0.33 0.15 0.11 1.06
## AU20_r_mean 87 235 0.07 0.04 0.06 0.06 0.02 0.02 0.32
## AU23_r_mean 88 235 0.08 0.05 0.06 0.07 0.04 0.01 0.32
## AU25_r_mean 89 235 0.46 0.31 0.37 0.42 0.27 0.08 1.70
## AU26_r_mean 90 235 0.31 0.17 0.27 0.29 0.13 0.09 1.41
## AU45_r_mean 91 235 0.16 0.11 0.13 0.15 0.08 0.04 0.63
## lip_mean 92 235 174.40 19.69 175.04 174.72 22.09 118.36 229.63
## eye_mean 93 235 10.29 2.18 10.24 10.23 2.15 5.32 17.31
## amplitude_mean 94 235 0.53 0.13 0.53 0.53 0.15 0.16 0.90
## stage_mean 95 235 0.04 0.02 0.04 0.04 0.01 0.02 0.13
## apex_mean 96 235 0.58 0.14 0.59 0.59 0.15 0.18 0.96
## offset_mean 97 235 0.46 0.13 0.46 0.46 0.14 0.14 0.73
## onset_mean 98 235 0.44 0.12 0.45 0.44 0.14 0.08 0.76
## frame_sd 99 235 70.78 30.48 64.81 67.77 26.96 16.02 203.66
## timestamp_sd 100 235 1.42 0.61 1.30 1.36 0.54 0.32 4.07
## gaze_angle_x_sd 101 235 0.02 0.01 0.02 0.02 0.01 0.00 0.10
## gaze_angle_y_sd 102 235 0.04 0.02 0.04 0.04 0.02 0.01 0.10
## pose_Rx_sd 103 235 0.03 0.02 0.02 0.03 0.01 0.01 0.16
## pose_Ry_sd 104 235 0.02 0.01 0.01 0.02 0.01 0.00 0.09
## pose_Rz_sd 105 235 0.03 0.02 0.02 0.02 0.01 0.00 0.17
## AU01_r_sd 106 235 0.18 0.16 0.14 0.15 0.08 0.04 1.60
## AU02_r_sd 107 235 0.12 0.09 0.09 0.10 0.05 0.03 0.62
## AU04_r_sd 108 235 0.12 0.16 0.05 0.09 0.07 0.00 0.67
## AU05_r_sd 109 235 0.10 0.05 0.08 0.09 0.04 0.03 0.35
## AU06_r_sd 110 235 0.55 0.21 0.54 0.54 0.18 0.00 1.15
## AU07_r_sd 111 235 0.42 0.22 0.41 0.41 0.18 0.00 1.26
## AU09_r_sd 112 235 0.09 0.07 0.07 0.08 0.04 0.02 0.56
## AU10_r_sd 113 235 0.40 0.22 0.41 0.40 0.20 0.00 1.15
## AU12_r_sd 114 235 0.74 0.19 0.72 0.73 0.21 0.32 1.34
## AU14_r_sd 115 235 0.30 0.12 0.27 0.29 0.11 0.01 0.83
## AU15_r_sd 116 235 0.11 0.07 0.10 0.11 0.04 0.04 0.64
## AU17_r_sd 117 235 0.35 0.18 0.32 0.33 0.16 0.08 1.23
## AU20_r_sd 118 235 0.12 0.06 0.10 0.11 0.05 0.03 0.52
## AU23_r_sd 119 235 0.15 0.09 0.13 0.13 0.08 0.03 0.54
## AU25_r_sd 120 235 0.42 0.28 0.33 0.39 0.23 0.07 1.43
## AU26_r_sd 121 235 0.33 0.18 0.29 0.31 0.14 0.07 1.46
## AU45_r_sd 122 235 0.33 0.23 0.28 0.30 0.22 0.05 1.21
## lip_sd 123 235 11.00 3.42 10.82 10.88 3.44 3.98 19.10
## eye_sd 124 235 1.38 0.69 1.27 1.33 0.70 0.23 3.77
## amplitude_sd 125 235 0.07 0.02 0.07 0.07 0.02 0.03 0.13
## stage_sd 126 235 0.01 0.01 0.01 0.01 0.00 0.00 0.10
## apex_sd 127 235 0.02 0.01 0.02 0.02 0.01 0.01 0.05
## offset_sd 128 235 0.04 0.02 0.03 0.04 0.02 0.00 0.14
## onset_sd 129 235 0.04 0.02 0.04 0.04 0.02 0.01 0.12
## range skew kurtosis se
## filename* 234.00 0.00 -1.22 4.43
## subject 514.00 0.19 -1.28 9.69
## gender* 1.00 -0.08 -2.00 0.03
## age 9.00 1.00 0.33 0.15
## smile_type* 0.00 NaN NaN 0.00
## frame_min 0.00 NaN NaN 0.00
## timestamp_min 0.00 NaN NaN 0.00
## gaze_angle_x_min 0.52 -0.61 1.46 0.01
## gaze_angle_y_min 0.61 -0.14 1.38 0.01
## pose_Rx_min 0.68 -0.71 0.86 0.01
## pose_Ry_min 0.50 -0.07 0.14 0.01
## pose_Rz_min 0.50 -0.68 1.20 0.01
## AU01_r_min 0.00 NaN NaN 0.00
## AU02_r_min 0.00 NaN NaN 0.00
## AU04_r_min 3.36 4.28 20.21 0.03
## AU05_r_min 0.00 NaN NaN 0.00
## AU06_r_min 1.82 4.11 20.83 0.01
## AU07_r_min 3.59 3.15 11.80 0.03
## AU09_r_min 0.00 NaN NaN 0.00
## AU10_r_min 0.86 4.29 18.95 0.01
## AU12_r_min 1.95 1.58 1.74 0.03
## AU14_r_min 1.85 0.47 -0.65 0.03
## AU15_r_min 0.00 NaN NaN 0.00
## AU17_r_min 0.00 NaN NaN 0.00
## AU20_r_min 0.00 NaN NaN 0.00
## AU23_r_min 0.00 NaN NaN 0.00
## AU25_r_min 0.00 NaN NaN 0.00
## AU26_r_min 0.00 NaN NaN 0.00
## AU45_r_min 0.00 NaN NaN 0.00
## lip_min 88.74 0.04 -0.40 1.16
## eye_min 15.10 0.57 -0.68 0.23
## amplitude_min 0.59 0.04 -0.40 0.01
## stage_min 0.07 2.43 9.81 0.00
## apex_min 0.75 -0.16 -0.32 0.01
## offset_min 0.55 -0.03 -0.68 0.01
## onset_min 0.70 0.07 -0.15 0.01
## frame_max 650.00 1.17 1.98 6.89
## timestamp_max 13.00 1.17 1.98 0.14
## gaze_angle_x_max 0.44 -0.05 0.57 0.00
## gaze_angle_y_max 0.55 -0.22 -0.39 0.01
## pose_Rx_max 0.51 -0.48 0.25 0.01
## pose_Ry_max 0.51 -0.02 -0.13 0.01
## pose_Rz_max 0.42 0.49 0.38 0.01
## AU01_r_max 4.37 2.92 13.47 0.03
## AU02_r_max 2.08 2.02 4.19 0.03
## AU04_r_max 4.49 1.81 3.22 0.06
## AU05_r_max 1.89 1.86 5.70 0.02
## AU06_r_max 3.74 0.10 -0.16 0.05
## AU07_r_max 5.00 0.50 0.45 0.06
## AU09_r_max 1.80 2.43 8.56 0.02
## AU10_r_max 3.34 0.04 -0.35 0.05
## AU12_r_max 3.01 -0.33 -0.07 0.04
## AU14_r_max 3.40 -0.27 1.13 0.03
## AU15_r_max 2.62 3.53 23.08 0.02
## AU17_r_max 3.66 1.17 2.29 0.04
## AU20_r_max 1.49 1.27 2.14 0.02
## AU23_r_max 1.81 1.19 1.27 0.02
## AU25_r_max 3.14 0.91 0.16 0.05
## AU26_r_max 4.44 1.97 6.50 0.04
## AU45_r_max 4.05 0.31 -1.24 0.07
## lip_max 117.59 -0.25 -0.24 1.36
## eye_max 14.80 0.38 1.12 0.15
## amplitude_max 0.78 -0.25 -0.24 0.01
## stage_max 0.45 6.04 56.02 0.00
## apex_max 0.78 -0.25 -0.24 0.01
## offset_max 0.74 -0.16 -0.35 0.01
## onset_max 0.73 -0.17 -0.35 0.01
## frame_mean 325.00 1.17 1.98 3.44
## timestamp_mean 6.50 1.17 1.98 0.07
## gaze_angle_x_mean 0.44 -0.04 0.53 0.00
## gaze_angle_y_mean 0.51 -0.20 1.00 0.00
## pose_Rx_mean 0.48 -0.47 0.11 0.01
## pose_Ry_mean 0.48 -0.08 0.07 0.01
## pose_Rz_mean 0.36 0.10 -0.41 0.00
## AU01_r_mean 0.92 4.32 29.72 0.01
## AU02_r_mean 0.27 3.07 10.93 0.00
## AU04_r_mean 3.75 2.80 8.51 0.04
## AU05_r_mean 0.11 1.66 3.35 0.00
## AU06_r_mean 2.95 0.56 0.02 0.04
## AU07_r_mean 4.76 1.11 1.82 0.05
## AU09_r_mean 0.29 3.55 17.18 0.00
## AU10_r_mean 2.13 0.40 -0.74 0.03
## AU12_r_mean 2.99 -0.09 -0.15 0.04
## AU14_r_mean 2.80 -0.19 0.44 0.03
## AU15_r_mean 0.26 2.04 5.97 0.00
## AU17_r_mean 0.94 1.29 1.67 0.01
## AU20_r_mean 0.30 2.62 11.87 0.00
## AU23_r_mean 0.30 1.57 3.25 0.00
## AU25_r_mean 1.61 1.28 1.67 0.02
## AU26_r_mean 1.32 2.37 9.87 0.01
## AU45_r_mean 0.59 1.59 2.74 0.01
## lip_mean 111.27 -0.17 -0.33 1.28
## eye_mean 11.98 0.38 0.35 0.14
## amplitude_mean 0.74 -0.17 -0.33 0.01
## stage_mean 0.11 2.76 11.95 0.00
## apex_mean 0.77 -0.21 -0.28 0.01
## offset_mean 0.60 -0.10 -0.57 0.01
## onset_mean 0.68 -0.12 -0.32 0.01
## frame_sd 187.64 1.17 1.98 1.99
## timestamp_sd 3.75 1.17 1.98 0.04
## gaze_angle_x_sd 0.09 2.71 11.41 0.00
## gaze_angle_y_sd 0.09 0.91 0.93 0.00
## pose_Rx_sd 0.15 2.20 6.91 0.00
## pose_Ry_sd 0.09 2.12 6.17 0.00
## pose_Rz_sd 0.16 2.74 10.42 0.00
## AU01_r_sd 1.56 4.42 31.16 0.01
## AU02_r_sd 0.59 2.66 8.05 0.01
## AU04_r_sd 0.67 1.57 2.06 0.01
## AU05_r_sd 0.32 1.86 4.41 0.00
## AU06_r_sd 1.15 0.10 0.33 0.01
## AU07_r_sd 1.26 0.66 1.12 0.01
## AU09_r_sd 0.54 3.27 15.18 0.00
## AU10_r_sd 1.15 0.19 0.26 0.01
## AU12_r_sd 1.02 0.29 -0.34 0.01
## AU14_r_sd 0.82 0.85 1.04 0.01
## AU15_r_sd 0.60 3.06 17.11 0.00
## AU17_r_sd 1.14 1.49 3.38 0.01
## AU20_r_sd 0.49 2.10 7.71 0.00
## AU23_r_sd 0.52 1.27 1.73 0.01
## AU25_r_sd 1.36 1.02 0.35 0.02
## AU26_r_sd 1.39 2.05 7.43 0.01
## AU45_r_sd 1.16 1.02 0.72 0.01
## lip_sd 15.12 0.31 -0.51 0.22
## eye_sd 3.54 0.75 0.46 0.05
## amplitude_sd 0.10 0.31 -0.51 0.00
## stage_sd 0.09 7.06 73.80 0.00
## apex_sd 0.05 0.74 0.37 0.00
## offset_sd 0.14 1.41 2.75 0.00
## onset_sd 0.11 0.85 0.59 0.00
# descriptive statistics per classifier and boys/girls
desc_stats_class_gender <- describeBy(UvA_sum ~ smile_type + gender)
desc_stats_class_gender
##
## Descriptive statistics by group
## smile_type: deliberate
## gender: female
## vars n mean sd median trimmed mad min max
## filename* 1 116 58.50 33.63 58.50 58.50 43.00 1.00 116.00
## subject 2 116 282.35 158.93 252.00 279.59 189.77 20.00 543.00
## gender* 3 116 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## age 4 116 11.46 2.51 11.00 11.28 2.97 8.00 17.00
## smile_type* 5 116 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## frame_min 6 116 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## timestamp_min 7 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## gaze_angle_x_min 8 116 0.14 0.08 0.15 0.14 0.07 -0.17 0.26
## gaze_angle_y_min 9 116 0.21 0.08 0.22 0.21 0.08 -0.01 0.48
## pose_Rx_min 10 116 0.13 0.09 0.14 0.13 0.10 -0.10 0.29
## pose_Ry_min 11 116 -0.18 0.07 -0.20 -0.18 0.08 -0.39 0.03
## pose_Rz_min 12 116 -0.04 0.06 -0.03 -0.03 0.06 -0.22 0.09
## AU01_r_min 13 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU02_r_min 14 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU04_r_min 15 116 0.16 0.54 0.00 0.02 0.00 0.00 2.99
## AU05_r_min 16 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU06_r_min 17 116 0.06 0.18 0.00 0.01 0.00 0.00 1.35
## AU07_r_min 18 116 0.27 0.56 0.00 0.14 0.00 0.00 3.46
## AU09_r_min 19 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU10_r_min 20 116 0.03 0.19 0.00 0.00 0.00 0.00 1.64
## AU12_r_min 21 116 0.24 0.38 0.00 0.16 0.00 0.00 2.56
## AU14_r_min 22 116 0.57 0.39 0.57 0.55 0.42 0.00 1.44
## AU15_r_min 23 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU17_r_min 24 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU20_r_min 25 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU23_r_min 26 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU25_r_min 27 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU26_r_min 28 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU45_r_min 29 116 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## lip_min 30 116 150.76 19.36 148.81 150.74 20.95 95.03 204.53
## eye_min 31 116 5.80 3.29 5.31 5.68 4.16 0.54 14.01
## amplitude_min 32 116 0.37 0.13 0.36 0.37 0.14 0.00 0.73
## stage_min 33 116 0.03 0.01 0.03 0.03 0.01 0.02 0.09
## apex_min 34 116 0.56 0.14 0.54 0.56 0.14 0.13 0.84
## offset_min 35 116 0.40 0.14 0.39 0.39 0.15 0.00 0.73
## onset_min 36 116 0.40 0.13 0.39 0.40 0.12 0.03 0.75
## frame_max 37 116 155.22 49.30 151.00 148.67 35.58 84.00 438.00
## timestamp_max 38 116 3.08 0.99 3.00 2.95 0.71 1.66 8.74
## gaze_angle_x_max 39 116 0.24 0.06 0.24 0.24 0.07 0.11 0.39
## gaze_angle_y_max 40 116 0.42 0.10 0.41 0.41 0.10 0.19 0.68
## pose_Rx_max 41 116 0.22 0.09 0.21 0.22 0.09 0.00 0.42
## pose_Ry_max 42 116 -0.12 0.07 -0.13 -0.12 0.08 -0.27 0.18
## pose_Rz_max 43 116 0.05 0.08 0.04 0.05 0.07 -0.14 0.33
## AU01_r_max 44 116 0.81 0.44 0.73 0.76 0.40 0.20 2.85
## AU02_r_max 45 116 0.50 0.25 0.47 0.47 0.21 0.13 1.60
## AU04_r_max 46 116 0.55 0.82 0.27 0.38 0.40 0.00 4.10
## AU05_r_max 47 116 0.52 0.35 0.42 0.47 0.27 0.10 2.34
## AU06_r_max 48 116 1.94 0.68 1.95 1.92 0.77 0.49 3.49
## AU07_r_max 49 116 1.85 0.93 1.83 1.83 0.81 0.01 5.00
## AU09_r_max 50 116 0.31 0.15 0.28 0.29 0.13 0.08 0.77
## AU10_r_max 51 116 1.31 0.73 1.33 1.31 0.76 0.00 2.86
## AU12_r_max 52 116 2.89 0.59 2.90 2.89 0.60 1.40 4.12
## AU14_r_max 53 116 1.86 0.39 1.84 1.87 0.43 1.02 2.70
## AU15_r_max 54 116 0.42 0.23 0.35 0.39 0.15 0.15 1.27
## AU17_r_max 55 116 1.33 0.64 1.27 1.25 0.49 0.29 3.72
## AU20_r_max 56 116 0.48 0.24 0.42 0.45 0.19 0.16 1.24
## AU23_r_max 57 116 0.59 0.36 0.54 0.55 0.36 0.09 1.88
## AU25_r_max 58 116 1.49 0.83 1.33 1.43 0.90 0.31 3.38
## AU26_r_max 59 116 1.12 0.48 1.02 1.07 0.44 0.31 2.86
## AU45_r_max 60 116 1.89 1.28 1.84 1.83 1.77 0.25 4.73
## lip_max 61 116 191.46 21.38 191.42 192.86 22.34 126.71 232.87
## eye_max 62 116 12.86 2.03 12.64 12.81 1.65 7.67 18.57
## amplitude_max 63 116 0.64 0.14 0.64 0.65 0.15 0.21 0.92
## stage_max 64 116 Inf NaN 0.06 0.06 0.02 0.03 Inf
## apex_max 65 116 0.64 0.14 0.64 0.65 0.15 0.21 0.92
## offset_max 66 116 0.54 0.14 0.53 0.55 0.15 0.13 0.84
## onset_max 67 116 0.54 0.14 0.53 0.55 0.14 0.13 0.84
## frame_mean 68 116 78.11 24.65 76.00 74.84 17.79 42.50 219.50
## timestamp_mean 69 116 1.54 0.49 1.50 1.48 0.36 0.83 4.37
## gaze_angle_x_mean 70 116 0.19 0.06 0.20 0.19 0.06 0.06 0.30
## gaze_angle_y_mean 71 116 0.29 0.08 0.29 0.29 0.06 0.12 0.52
## pose_Rx_mean 72 116 0.18 0.09 0.18 0.18 0.10 -0.03 0.37
## pose_Ry_mean 73 116 -0.15 0.07 -0.17 -0.15 0.07 -0.32 0.05
## pose_Rz_mean 74 116 0.01 0.06 0.01 0.01 0.06 -0.16 0.24
## AU01_r_mean 75 116 0.14 0.08 0.13 0.13 0.08 0.03 0.45
## AU02_r_mean 76 116 0.05 0.03 0.05 0.05 0.02 0.02 0.20
## AU04_r_mean 77 116 0.30 0.66 0.02 0.15 0.03 0.00 3.50
## AU05_r_mean 78 116 0.05 0.03 0.04 0.04 0.02 0.01 0.24
## AU06_r_mean 79 116 1.14 0.58 1.08 1.11 0.65 0.09 2.82
## AU07_r_mean 80 116 1.08 0.84 0.98 1.00 0.78 0.00 4.89
## AU09_r_mean 81 116 0.04 0.02 0.03 0.04 0.02 0.01 0.13
## AU10_r_mean 82 116 0.70 0.51 0.72 0.67 0.56 0.00 2.06
## AU12_r_mean 83 116 1.97 0.56 1.93 1.96 0.61 0.78 3.34
## AU14_r_mean 84 116 1.34 0.37 1.31 1.35 0.40 0.46 2.13
## AU15_r_mean 85 116 0.07 0.04 0.06 0.07 0.03 0.03 0.24
## AU17_r_mean 86 116 0.40 0.22 0.35 0.36 0.14 0.11 1.67
## AU20_r_mean 87 116 0.07 0.04 0.06 0.07 0.03 0.02 0.27
## AU23_r_mean 88 116 0.09 0.06 0.07 0.08 0.04 0.02 0.28
## AU25_r_mean 89 116 0.67 0.46 0.59 0.63 0.52 0.08 1.89
## AU26_r_mean 90 116 0.29 0.15 0.26 0.27 0.11 0.10 1.13
## AU45_r_mean 91 116 0.20 0.14 0.18 0.19 0.14 0.04 0.96
## lip_mean 92 116 177.74 20.65 175.98 178.86 21.06 115.14 220.98
## eye_mean 93 116 10.56 1.85 10.54 10.61 1.40 4.65 16.01
## amplitude_mean 94 116 0.55 0.14 0.54 0.56 0.14 0.13 0.84
## stage_mean 95 116 Inf NaN 0.04 0.04 0.01 0.02 Inf
## apex_mean 96 116 0.62 0.14 0.61 0.63 0.15 0.19 0.89
## offset_mean 97 116 0.45 0.14 0.43 0.45 0.15 0.03 0.78
## onset_mean 98 116 0.44 0.13 0.44 0.45 0.13 0.04 0.78
## frame_sd 99 116 44.95 14.23 43.73 43.06 10.27 24.39 126.58
## timestamp_sd 100 116 0.90 0.28 0.87 0.86 0.21 0.49 2.53
## gaze_angle_x_sd 101 116 0.02 0.02 0.02 0.02 0.01 0.01 0.10
## gaze_angle_y_sd 102 116 0.04 0.02 0.04 0.04 0.02 0.01 0.12
## pose_Rx_sd 103 116 0.02 0.01 0.02 0.02 0.01 0.01 0.07
## pose_Ry_sd 104 116 0.02 0.01 0.01 0.01 0.01 0.00 0.07
## pose_Rz_sd 105 116 0.03 0.03 0.02 0.02 0.01 0.00 0.14
## AU01_r_sd 106 116 0.22 0.12 0.20 0.20 0.13 0.04 0.80
## AU02_r_sd 107 116 0.11 0.07 0.10 0.10 0.05 0.03 0.44
## AU04_r_sd 108 116 0.10 0.12 0.04 0.07 0.07 0.00 0.59
## AU05_r_sd 109 116 0.11 0.08 0.08 0.10 0.05 0.03 0.61
## AU06_r_sd 110 116 0.66 0.23 0.66 0.65 0.27 0.14 1.31
## AU07_r_sd 111 116 0.44 0.22 0.43 0.44 0.19 0.00 1.02
## AU09_r_sd 112 116 0.08 0.04 0.07 0.07 0.04 0.02 0.22
## AU10_r_sd 113 116 0.45 0.27 0.44 0.45 0.30 0.00 1.12
## AU12_r_sd 114 116 0.91 0.26 0.88 0.90 0.28 0.41 1.61
## AU14_r_sd 115 116 0.33 0.12 0.31 0.32 0.13 0.13 0.72
## AU15_r_sd 116 116 0.11 0.06 0.09 0.10 0.04 0.04 0.38
## AU17_r_sd 117 116 0.42 0.24 0.38 0.39 0.15 0.08 1.49
## AU20_r_sd 118 116 0.13 0.08 0.10 0.11 0.06 0.04 0.42
## AU23_r_sd 119 116 0.16 0.10 0.13 0.14 0.09 0.02 0.45
## AU25_r_sd 120 116 0.55 0.37 0.46 0.52 0.36 0.08 1.44
## AU26_r_sd 121 116 0.33 0.17 0.29 0.31 0.13 0.09 1.08
## AU45_r_sd 122 116 0.41 0.29 0.35 0.39 0.33 0.06 1.50
## lip_sd 123 116 13.43 3.81 13.54 13.45 4.07 5.64 21.98
## eye_sd 124 116 1.52 0.81 1.37 1.45 0.83 0.20 4.24
## amplitude_sd 125 116 0.09 0.03 0.09 0.09 0.03 0.04 0.15
## stage_sd 126 115 0.01 0.02 0.01 0.01 0.00 0.00 0.16
## apex_sd 127 116 0.02 0.01 0.02 0.02 0.01 0.01 0.04
## offset_sd 128 116 0.04 0.02 0.04 0.04 0.02 0.01 0.10
## onset_sd 129 116 0.05 0.02 0.05 0.05 0.02 0.00 0.10
## range skew kurtosis se
## filename* 115.00 0.00 -1.23 3.12
## subject 523.00 0.22 -1.29 14.76
## gender* 0.00 NaN NaN 0.00
## age 9.00 0.58 -0.58 0.23
## smile_type* 0.00 NaN NaN 0.00
## frame_min 0.00 NaN NaN 0.00
## timestamp_min 0.00 NaN NaN 0.00
## gaze_angle_x_min 0.44 -1.20 2.05 0.01
## gaze_angle_y_min 0.49 0.15 0.78 0.01
## pose_Rx_min 0.39 -0.25 -0.67 0.01
## pose_Ry_min 0.42 0.21 -0.03 0.01
## pose_Rz_min 0.32 -0.45 -0.09 0.01
## AU01_r_min 0.00 NaN NaN 0.00
## AU02_r_min 0.00 NaN NaN 0.00
## AU04_r_min 2.99 4.21 17.22 0.05
## AU05_r_min 0.00 NaN NaN 0.00
## AU06_r_min 1.35 4.47 24.86 0.02
## AU07_r_min 3.46 2.83 10.14 0.05
## AU09_r_min 0.00 NaN NaN 0.00
## AU10_r_min 1.64 6.85 49.77 0.02
## AU12_r_min 2.56 2.70 11.12 0.04
## AU14_r_min 1.44 0.14 -0.86 0.04
## AU15_r_min 0.00 NaN NaN 0.00
## AU17_r_min 0.00 NaN NaN 0.00
## AU20_r_min 0.00 NaN NaN 0.00
## AU23_r_min 0.00 NaN NaN 0.00
## AU25_r_min 0.00 NaN NaN 0.00
## AU26_r_min 0.00 NaN NaN 0.00
## AU45_r_min 0.00 NaN NaN 0.00
## lip_min 109.50 0.04 0.11 1.80
## eye_min 13.47 0.30 -1.12 0.31
## amplitude_min 0.73 0.04 0.11 0.01
## stage_min 0.07 2.81 11.05 0.00
## apex_min 0.71 -0.48 -0.03 0.01
## offset_min 0.73 0.00 -0.13 0.01
## onset_min 0.73 -0.16 0.34 0.01
## frame_max 354.00 2.47 9.87 4.58
## timestamp_max 7.08 2.47 9.87 0.09
## gaze_angle_x_max 0.28 -0.02 -0.60 0.01
## gaze_angle_y_max 0.49 0.28 -0.45 0.01
## pose_Rx_max 0.42 0.03 -0.54 0.01
## pose_Ry_max 0.45 0.77 1.36 0.01
## pose_Rz_max 0.48 1.03 2.06 0.01
## AU01_r_max 2.65 1.32 2.96 0.04
## AU02_r_max 1.47 1.77 4.61 0.02
## AU04_r_max 4.10 2.48 6.79 0.08
## AU05_r_max 2.24 1.91 5.75 0.03
## AU06_r_max 3.00 0.21 -0.60 0.06
## AU07_r_max 4.99 0.46 1.02 0.09
## AU09_r_max 0.69 0.89 0.18 0.01
## AU10_r_max 2.86 -0.10 -0.76 0.07
## AU12_r_max 2.72 -0.07 -0.40 0.05
## AU14_r_max 1.68 -0.06 -0.73 0.04
## AU15_r_max 1.12 1.54 2.38 0.02
## AU17_r_max 3.43 1.38 2.67 0.06
## AU20_r_max 1.08 1.10 0.74 0.02
## AU23_r_max 1.79 1.10 0.91 0.03
## AU25_r_max 3.07 0.51 -0.79 0.08
## AU26_r_max 2.55 0.96 0.93 0.04
## AU45_r_max 4.48 0.22 -1.33 0.12
## lip_max 106.16 -0.54 0.11 1.98
## eye_max 10.90 0.24 0.28 0.19
## amplitude_max 0.71 -0.54 0.11 0.01
## stage_max Inf NaN NaN NaN
## apex_max 0.71 -0.54 0.11 0.01
## offset_max 0.71 -0.43 -0.02 0.01
## onset_max 0.70 -0.47 0.00 0.01
## frame_mean 177.00 2.47 9.87 2.29
## timestamp_mean 3.54 2.47 9.87 0.05
## gaze_angle_x_mean 0.24 -0.26 -0.76 0.01
## gaze_angle_y_mean 0.40 0.52 0.51 0.01
## pose_Rx_mean 0.40 -0.05 -0.60 0.01
## pose_Ry_mean 0.36 0.34 -0.09 0.01
## pose_Rz_mean 0.40 0.28 1.15 0.01
## AU01_r_mean 0.42 1.26 2.26 0.01
## AU02_r_mean 0.18 2.56 8.08 0.00
## AU04_r_mean 3.50 3.35 11.65 0.06
## AU05_r_mean 0.22 2.54 10.89 0.00
## AU06_r_mean 2.73 0.49 -0.41 0.05
## AU07_r_mean 4.89 1.44 3.69 0.08
## AU09_r_mean 0.12 1.38 1.98 0.00
## AU10_r_mean 2.06 0.52 -0.40 0.05
## AU12_r_mean 2.56 0.10 -0.39 0.05
## AU14_r_mean 1.67 -0.06 -0.64 0.03
## AU15_r_mean 0.21 1.91 4.99 0.00
## AU17_r_mean 1.56 2.47 9.80 0.02
## AU20_r_mean 0.25 1.83 3.92 0.00
## AU23_r_mean 0.27 1.30 1.22 0.01
## AU25_r_mean 1.81 0.65 -0.54 0.04
## AU26_r_mean 1.03 2.32 8.85 0.01
## AU45_r_mean 0.91 1.80 5.69 0.01
## lip_mean 105.84 -0.47 -0.04 1.92
## eye_mean 11.36 -0.20 0.83 0.17
## amplitude_mean 0.71 -0.47 -0.04 0.01
## stage_mean Inf NaN NaN NaN
## apex_mean 0.70 -0.52 -0.08 0.01
## offset_mean 0.75 -0.15 0.09 0.01
## onset_mean 0.74 -0.33 0.40 0.01
## frame_sd 102.19 2.47 9.87 1.32
## timestamp_sd 2.04 2.47 9.87 0.03
## gaze_angle_x_sd 0.10 2.43 7.90 0.00
## gaze_angle_y_sd 0.11 0.83 1.18 0.00
## pose_Rx_sd 0.06 1.16 1.04 0.00
## pose_Ry_sd 0.07 2.04 4.45 0.00
## pose_Rz_sd 0.14 2.25 5.42 0.00
## AU01_r_sd 0.76 1.36 3.47 0.01
## AU02_r_sd 0.41 2.51 7.97 0.01
## AU04_r_sd 0.59 1.68 2.93 0.01
## AU05_r_sd 0.58 2.46 10.38 0.01
## AU06_r_sd 1.18 0.23 -0.50 0.02
## AU07_r_sd 1.02 0.19 -0.03 0.02
## AU09_r_sd 0.21 1.17 1.09 0.00
## AU10_r_sd 1.12 0.02 -0.74 0.02
## AU12_r_sd 1.21 0.60 -0.23 0.02
## AU14_r_sd 0.58 0.75 0.33 0.01
## AU15_r_sd 0.34 1.90 4.48 0.01
## AU17_r_sd 1.41 1.89 5.13 0.02
## AU20_r_sd 0.38 1.55 2.10 0.01
## AU23_r_sd 0.43 1.10 0.48 0.01
## AU25_r_sd 1.36 0.63 -0.68 0.03
## AU26_r_sd 0.99 1.55 3.67 0.02
## AU45_r_sd 1.44 0.81 0.47 0.03
## lip_sd 16.35 -0.01 -0.77 0.35
## eye_sd 4.03 0.77 0.24 0.08
## amplitude_sd 0.11 -0.01 -0.77 0.00
## stage_sd 0.16 8.92 86.69 0.00
## apex_sd 0.04 0.37 -0.42 0.00
## offset_sd 0.09 0.42 0.19 0.00
## onset_sd 0.10 0.30 -0.17 0.00
## ------------------------------------------------------------
## smile_type: spontaneous
## gender: female
## vars n mean sd median trimmed mad min max
## filename* 1 113 57.00 32.76 57.00 57.00 41.51 1.00 113.00
## subject 2 113 275.07 154.03 251.00 271.66 185.32 20.00 534.00
## gender* 3 113 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## age 4 113 11.01 2.45 10.00 10.75 2.97 8.00 17.00
## smile_type* 5 113 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## frame_min 6 113 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## timestamp_min 7 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## gaze_angle_x_min 8 113 0.15 0.07 0.16 0.16 0.06 -0.13 0.28
## gaze_angle_y_min 9 113 0.20 0.08 0.20 0.20 0.08 -0.07 0.48
## pose_Rx_min 10 113 0.10 0.11 0.12 0.11 0.08 -0.37 0.31
## pose_Ry_min 11 113 -0.20 0.07 -0.20 -0.20 0.06 -0.38 0.04
## pose_Rz_min 12 113 -0.05 0.09 -0.04 -0.04 0.08 -0.34 0.12
## AU01_r_min 13 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU02_r_min 14 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU04_r_min 15 113 0.14 0.53 0.00 0.01 0.00 0.00 3.36
## AU05_r_min 16 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU06_r_min 17 113 0.07 0.18 0.00 0.02 0.00 0.00 0.93
## AU07_r_min 18 113 0.28 0.56 0.00 0.15 0.00 0.00 3.59
## AU09_r_min 19 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU10_r_min 20 113 0.03 0.12 0.00 0.00 0.00 0.00 0.86
## AU12_r_min 21 113 0.44 0.49 0.28 0.37 0.42 0.00 1.76
## AU14_r_min 22 113 0.64 0.46 0.59 0.62 0.52 0.00 1.59
## AU15_r_min 23 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU17_r_min 24 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU20_r_min 25 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU23_r_min 26 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU25_r_min 27 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU26_r_min 28 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU45_r_min 29 113 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## lip_min 30 113 151.98 18.28 151.90 152.23 17.85 101.23 189.97
## eye_min 31 113 5.72 3.43 4.70 5.51 3.93 0.84 13.82
## amplitude_min 32 113 0.38 0.12 0.38 0.38 0.12 0.04 0.63
## stage_min 33 113 0.03 0.01 0.03 0.03 0.01 0.02 0.09
## apex_min 34 113 0.55 0.14 0.55 0.55 0.15 0.16 0.80
## offset_min 35 113 0.42 0.13 0.43 0.43 0.13 0.12 0.65
## onset_min 36 113 0.40 0.12 0.40 0.40 0.12 0.04 0.68
## frame_max 37 113 237.58 100.18 212.00 226.31 71.16 77.00 606.00
## timestamp_max 38 113 4.73 2.00 4.22 4.51 1.42 1.52 12.10
## gaze_angle_x_max 39 113 0.25 0.06 0.25 0.25 0.06 0.11 0.38
## gaze_angle_y_max 40 113 0.42 0.10 0.42 0.42 0.12 0.14 0.62
## pose_Rx_max 41 113 0.23 0.08 0.22 0.23 0.08 0.00 0.42
## pose_Ry_max 42 113 -0.12 0.08 -0.13 -0.12 0.08 -0.30 0.08
## pose_Rz_max 43 113 0.06 0.08 0.05 0.05 0.07 -0.10 0.30
## AU01_r_max 44 113 0.70 0.39 0.61 0.65 0.30 0.15 2.30
## AU02_r_max 45 113 0.58 0.39 0.44 0.50 0.24 0.17 2.24
## AU04_r_max 46 113 0.66 0.86 0.40 0.49 0.59 0.00 4.49
## AU05_r_max 47 113 0.52 0.29 0.46 0.48 0.24 0.12 2.01
## AU06_r_max 48 113 2.10 0.70 2.03 2.08 0.70 0.44 3.64
## AU07_r_max 49 113 2.05 0.94 2.04 2.04 0.86 0.00 5.00
## AU09_r_max 50 113 0.38 0.25 0.30 0.34 0.15 0.08 1.88
## AU10_r_max 51 113 1.55 0.74 1.61 1.53 0.76 0.00 3.34
## AU12_r_max 52 113 2.89 0.56 2.96 2.92 0.59 1.19 4.20
## AU14_r_max 53 113 1.91 0.39 1.91 1.91 0.43 1.03 2.91
## AU15_r_max 54 113 0.46 0.24 0.42 0.43 0.21 0.14 1.77
## AU17_r_max 55 113 1.39 0.64 1.35 1.33 0.44 0.30 3.96
## AU20_r_max 56 113 0.52 0.28 0.44 0.48 0.24 0.15 1.61
## AU23_r_max 57 113 0.64 0.38 0.58 0.59 0.37 0.11 1.91
## AU25_r_max 58 113 1.31 0.71 1.15 1.24 0.65 0.32 3.12
## AU26_r_max 59 113 1.16 0.51 1.09 1.10 0.39 0.31 3.27
## AU45_r_max 60 113 1.64 1.07 1.53 1.58 1.44 0.22 4.27
## lip_max 61 113 190.36 21.24 190.20 191.59 19.92 127.31 228.96
## eye_max 62 113 12.63 2.22 12.75 12.65 2.07 7.46 18.17
## amplitude_max 63 113 0.64 0.14 0.64 0.64 0.13 0.22 0.89
## stage_max 64 113 0.06 0.05 0.05 0.06 0.02 0.03 0.48
## apex_max 65 113 0.64 0.14 0.64 0.64 0.13 0.22 0.89
## offset_max 66 113 0.54 0.14 0.55 0.55 0.15 0.15 0.80
## onset_max 67 113 0.54 0.13 0.55 0.55 0.14 0.16 0.79
## frame_mean 68 113 119.29 50.09 106.50 113.65 35.58 39.00 303.50
## timestamp_mean 69 113 2.37 1.00 2.11 2.25 0.71 0.76 6.05
## gaze_angle_x_mean 70 113 0.20 0.06 0.20 0.20 0.05 0.07 0.32
## gaze_angle_y_mean 71 113 0.29 0.07 0.28 0.28 0.06 0.12 0.50
## pose_Rx_mean 72 113 0.17 0.08 0.17 0.18 0.07 -0.02 0.37
## pose_Ry_mean 73 113 -0.16 0.07 -0.17 -0.16 0.06 -0.35 0.05
## pose_Rz_mean 74 113 0.01 0.07 0.00 0.00 0.07 -0.16 0.16
## AU01_r_mean 75 113 0.11 0.06 0.10 0.10 0.05 0.03 0.36
## AU02_r_mean 76 113 0.05 0.04 0.04 0.05 0.02 0.02 0.29
## AU04_r_mean 77 113 0.31 0.65 0.03 0.16 0.04 0.00 3.75
## AU05_r_mean 78 113 0.04 0.02 0.03 0.04 0.01 0.01 0.12
## AU06_r_mean 79 113 1.24 0.56 1.22 1.22 0.63 0.10 2.89
## AU07_r_mean 80 113 1.18 0.83 1.15 1.12 0.88 0.00 4.76
## AU09_r_mean 81 113 0.05 0.04 0.03 0.04 0.01 0.01 0.29
## AU10_r_mean 82 113 0.84 0.52 0.88 0.83 0.62 0.00 2.13
## AU12_r_mean 83 113 2.05 0.55 2.05 2.07 0.55 0.41 3.40
## AU14_r_mean 84 113 1.36 0.37 1.37 1.37 0.43 0.53 2.21
## AU15_r_mean 85 113 0.08 0.04 0.06 0.07 0.03 0.03 0.23
## AU17_r_mean 86 113 0.39 0.21 0.35 0.37 0.18 0.11 1.06
## AU20_r_mean 87 113 0.07 0.04 0.06 0.06 0.03 0.02 0.32
## AU23_r_mean 88 113 0.08 0.05 0.07 0.07 0.04 0.02 0.32
## AU25_r_mean 89 113 0.48 0.33 0.39 0.44 0.30 0.08 1.63
## AU26_r_mean 90 113 0.30 0.15 0.26 0.28 0.12 0.09 0.80
## AU45_r_mean 91 113 0.17 0.12 0.13 0.15 0.09 0.04 0.63
## lip_mean 92 113 176.97 20.27 177.87 177.92 21.92 118.36 214.74
## eye_mean 93 113 10.20 2.11 10.31 10.24 2.02 5.32 15.40
## amplitude_mean 94 113 0.55 0.14 0.55 0.55 0.15 0.16 0.80
## stage_mean 95 113 0.04 0.02 0.04 0.04 0.01 0.03 0.13
## apex_mean 96 113 0.60 0.14 0.60 0.61 0.14 0.18 0.87
## offset_mean 97 113 0.48 0.13 0.48 0.48 0.14 0.14 0.73
## onset_mean 98 113 0.46 0.13 0.46 0.46 0.14 0.08 0.72
## frame_sd 99 113 68.73 28.92 61.34 65.47 20.54 22.37 175.08
## timestamp_sd 100 113 1.37 0.58 1.23 1.31 0.41 0.45 3.50
## gaze_angle_x_sd 101 113 0.02 0.01 0.02 0.02 0.01 0.01 0.10
## gaze_angle_y_sd 102 113 0.04 0.02 0.04 0.04 0.02 0.01 0.10
## pose_Rx_sd 103 113 0.03 0.02 0.02 0.03 0.01 0.01 0.16
## pose_Ry_sd 104 113 0.02 0.01 0.01 0.02 0.01 0.00 0.09
## pose_Rz_sd 105 113 0.03 0.03 0.02 0.02 0.01 0.00 0.17
## AU01_r_sd 106 113 0.17 0.10 0.15 0.15 0.08 0.04 0.57
## AU02_r_sd 107 113 0.12 0.09 0.09 0.10 0.05 0.03 0.62
## AU04_r_sd 108 113 0.12 0.15 0.07 0.10 0.11 0.00 0.66
## AU05_r_sd 109 113 0.10 0.06 0.08 0.09 0.03 0.03 0.35
## AU06_r_sd 110 113 0.60 0.21 0.56 0.59 0.19 0.13 1.15
## AU07_r_sd 111 113 0.44 0.23 0.42 0.42 0.17 0.00 1.26
## AU09_r_sd 112 113 0.09 0.07 0.07 0.08 0.03 0.02 0.56
## AU10_r_sd 113 113 0.45 0.23 0.44 0.45 0.22 0.00 1.15
## AU12_r_sd 114 113 0.71 0.19 0.69 0.70 0.21 0.33 1.14
## AU14_r_sd 115 113 0.29 0.11 0.27 0.28 0.12 0.07 0.64
## AU15_r_sd 116 113 0.11 0.06 0.09 0.10 0.04 0.04 0.43
## AU17_r_sd 117 113 0.39 0.21 0.35 0.36 0.16 0.08 1.23
## AU20_r_sd 118 113 0.12 0.08 0.10 0.11 0.05 0.03 0.52
## AU23_r_sd 119 113 0.15 0.09 0.13 0.14 0.09 0.03 0.54
## AU25_r_sd 120 113 0.44 0.28 0.39 0.41 0.28 0.07 1.22
## AU26_r_sd 121 113 0.32 0.16 0.30 0.31 0.14 0.07 0.94
## AU45_r_sd 122 113 0.33 0.24 0.27 0.30 0.23 0.05 1.21
## lip_sd 123 113 10.86 3.41 10.53 10.75 3.43 3.98 18.73
## eye_sd 124 113 1.43 0.72 1.27 1.37 0.69 0.23 3.77
## amplitude_sd 125 113 0.07 0.02 0.07 0.07 0.02 0.03 0.12
## stage_sd 126 113 0.01 0.01 0.01 0.01 0.00 0.00 0.10
## apex_sd 127 113 0.02 0.01 0.02 0.02 0.01 0.01 0.05
## offset_sd 128 113 0.04 0.02 0.03 0.03 0.02 0.00 0.14
## onset_sd 129 113 0.05 0.02 0.04 0.04 0.02 0.01 0.12
## range skew kurtosis se
## filename* 112.00 0.00 -1.23 3.08
## subject 514.00 0.26 -1.23 14.49
## gender* 0.00 NaN NaN 0.00
## age 9.00 0.79 -0.11 0.23
## smile_type* 0.00 NaN NaN 0.00
## frame_min 0.00 NaN NaN 0.00
## timestamp_min 0.00 NaN NaN 0.00
## gaze_angle_x_min 0.41 -0.83 1.27 0.01
## gaze_angle_y_min 0.56 0.46 1.76 0.01
## pose_Rx_min 0.68 -0.96 2.32 0.01
## pose_Ry_min 0.41 0.32 0.30 0.01
## pose_Rz_min 0.46 -0.78 1.21 0.01
## AU01_r_min 0.00 NaN NaN 0.00
## AU02_r_min 0.00 NaN NaN 0.00
## AU04_r_min 3.36 4.45 20.10 0.05
## AU05_r_min 0.00 NaN NaN 0.00
## AU06_r_min 0.93 2.72 7.09 0.02
## AU07_r_min 3.59 3.20 12.95 0.05
## AU09_r_min 0.00 NaN NaN 0.00
## AU10_r_min 0.86 4.26 20.70 0.01
## AU12_r_min 1.76 0.89 -0.38 0.05
## AU14_r_min 1.59 0.29 -1.04 0.04
## AU15_r_min 0.00 NaN NaN 0.00
## AU17_r_min 0.00 NaN NaN 0.00
## AU20_r_min 0.00 NaN NaN 0.00
## AU23_r_min 0.00 NaN NaN 0.00
## AU25_r_min 0.00 NaN NaN 0.00
## AU26_r_min 0.00 NaN NaN 0.00
## AU45_r_min 0.00 NaN NaN 0.00
## lip_min 88.74 -0.16 -0.14 1.72
## eye_min 12.98 0.46 -1.04 0.32
## amplitude_min 0.59 -0.16 -0.14 0.01
## stage_min 0.07 3.04 11.51 0.00
## apex_min 0.65 -0.49 0.07 0.01
## offset_min 0.52 -0.27 -0.51 0.01
## onset_min 0.64 -0.19 -0.02 0.01
## frame_max 529.00 1.18 1.41 9.42
## timestamp_max 10.58 1.18 1.41 0.19
## gaze_angle_x_max 0.28 -0.40 -0.29 0.01
## gaze_angle_y_max 0.48 -0.09 -0.73 0.01
## pose_Rx_max 0.42 -0.37 0.30 0.01
## pose_Ry_max 0.38 0.14 -0.41 0.01
## pose_Rz_max 0.41 0.61 0.48 0.01
## AU01_r_max 2.15 1.27 1.93 0.04
## AU02_r_max 2.07 2.15 5.20 0.04
## AU04_r_max 4.49 2.07 4.93 0.08
## AU05_r_max 1.89 1.88 5.46 0.03
## AU06_r_max 3.20 0.24 -0.56 0.07
## AU07_r_max 5.00 0.32 0.60 0.09
## AU09_r_max 1.80 2.62 10.75 0.02
## AU10_r_max 3.34 0.12 -0.32 0.07
## AU12_r_max 3.01 -0.43 0.13 0.05
## AU14_r_max 1.88 0.00 -0.76 0.04
## AU15_r_max 1.63 2.10 7.59 0.02
## AU17_r_max 3.66 1.33 2.99 0.06
## AU20_r_max 1.46 1.36 2.10 0.03
## AU23_r_max 1.80 1.13 1.12 0.04
## AU25_r_max 2.80 0.78 -0.18 0.07
## AU26_r_max 2.96 1.32 2.77 0.05
## AU45_r_max 4.05 0.34 -1.16 0.10
## lip_max 101.65 -0.60 0.35 2.00
## eye_max 10.71 -0.08 -0.04 0.21
## amplitude_max 0.68 -0.60 0.35 0.01
## stage_max 0.45 6.08 46.18 0.00
## apex_max 0.68 -0.60 0.35 0.01
## offset_max 0.64 -0.47 0.04 0.01
## onset_max 0.64 -0.49 0.06 0.01
## frame_mean 264.50 1.18 1.41 4.71
## timestamp_mean 5.29 1.18 1.41 0.09
## gaze_angle_x_mean 0.26 -0.27 -0.21 0.01
## gaze_angle_y_mean 0.39 0.61 0.79 0.01
## pose_Rx_mean 0.39 -0.11 -0.18 0.01
## pose_Ry_mean 0.40 0.15 0.17 0.01
## pose_Rz_mean 0.32 0.10 -0.54 0.01
## AU01_r_mean 0.32 1.38 2.10 0.01
## AU02_r_mean 0.27 3.22 14.34 0.00
## AU04_r_mean 3.75 3.27 11.58 0.06
## AU05_r_mean 0.10 1.49 2.43 0.00
## AU06_r_mean 2.80 0.50 0.02 0.05
## AU07_r_mean 4.76 1.03 2.33 0.08
## AU09_r_mean 0.28 3.48 17.03 0.00
## AU10_r_mean 2.13 0.23 -0.75 0.05
## AU12_r_mean 2.98 -0.24 -0.04 0.05
## AU14_r_mean 1.68 -0.10 -0.90 0.04
## AU15_r_mean 0.20 1.62 2.73 0.00
## AU17_r_mean 0.94 1.07 0.90 0.02
## AU20_r_mean 0.29 2.65 9.82 0.00
## AU23_r_mean 0.30 1.79 4.18 0.00
## AU25_r_mean 1.55 1.13 0.91 0.03
## AU26_r_mean 0.71 1.25 1.56 0.01
## AU45_r_mean 0.59 1.60 2.45 0.01
## lip_mean 96.38 -0.49 0.05 1.91
## eye_mean 10.08 -0.12 -0.18 0.20
## amplitude_mean 0.64 -0.49 0.05 0.01
## stage_mean 0.11 3.25 12.67 0.00
## apex_mean 0.68 -0.57 0.24 0.01
## offset_mean 0.59 -0.29 -0.36 0.01
## onset_mean 0.64 -0.33 0.08 0.01
## frame_sd 152.71 1.18 1.41 2.72
## timestamp_sd 3.05 1.18 1.41 0.05
## gaze_angle_x_sd 0.09 2.94 14.68 0.00
## gaze_angle_y_sd 0.09 0.75 0.52 0.00
## pose_Rx_sd 0.15 2.52 7.95 0.00
## pose_Ry_sd 0.09 2.23 6.53 0.00
## pose_Rz_sd 0.16 2.52 7.85 0.00
## AU01_r_sd 0.53 1.27 1.70 0.01
## AU02_r_sd 0.59 2.88 10.91 0.01
## AU04_r_sd 0.66 1.56 2.43 0.01
## AU05_r_sd 0.32 1.80 3.81 0.01
## AU06_r_sd 1.03 0.52 -0.22 0.02
## AU07_r_sd 1.26 0.93 1.53 0.02
## AU09_r_sd 0.54 3.32 16.13 0.01
## AU10_r_sd 1.15 0.37 0.24 0.02
## AU12_r_sd 0.80 0.28 -0.62 0.02
## AU14_r_sd 0.57 0.71 0.07 0.01
## AU15_r_sd 0.39 1.98 5.75 0.01
## AU17_r_sd 1.14 1.42 2.61 0.02
## AU20_r_sd 0.49 2.19 6.86 0.01
## AU23_r_sd 0.51 1.39 2.27 0.01
## AU25_r_sd 1.15 0.85 -0.19 0.03
## AU26_r_sd 0.87 1.30 2.18 0.02
## AU45_r_sd 1.16 1.14 1.20 0.02
## lip_sd 14.75 0.27 -0.63 0.32
## eye_sd 3.54 0.89 0.77 0.07
## amplitude_sd 0.10 0.27 -0.63 0.00
## stage_sd 0.09 6.61 53.38 0.00
## apex_sd 0.05 0.89 0.65 0.00
## offset_sd 0.13 1.49 2.81 0.00
## onset_sd 0.11 0.90 0.83 0.00
## ------------------------------------------------------------
## smile_type: deliberate
## gender: male
## vars n mean sd median trimmed mad min max
## filename* 1 124 62.50 35.94 62.50 62.50 45.96 1.00 124.00
## subject 2 124 297.57 151.14 282.50 300.13 220.17 54.00 520.00
## gender* 3 124 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## age 4 124 10.49 2.11 10.00 10.21 1.48 8.00 17.00
## smile_type* 5 124 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## frame_min 6 124 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## timestamp_min 7 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## gaze_angle_x_min 8 124 0.15 0.09 0.16 0.15 0.07 -0.12 0.37
## gaze_angle_y_min 9 124 0.20 0.09 0.22 0.21 0.08 -0.10 0.38
## pose_Rx_min 10 124 0.10 0.11 0.11 0.11 0.10 -0.33 0.29
## pose_Ry_min 11 124 -0.20 0.09 -0.20 -0.20 0.07 -0.44 0.01
## pose_Rz_min 12 124 -0.04 0.07 -0.03 -0.04 0.07 -0.26 0.09
## AU01_r_min 13 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU02_r_min 14 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU04_r_min 15 124 0.12 0.40 0.00 0.01 0.00 0.00 2.18
## AU05_r_min 16 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU06_r_min 17 124 0.04 0.20 0.00 0.00 0.00 0.00 1.21
## AU07_r_min 18 124 0.12 0.40 0.00 0.01 0.00 0.00 2.36
## AU09_r_min 19 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU10_r_min 20 124 0.00 0.02 0.00 0.00 0.00 0.00 0.24
## AU12_r_min 21 124 0.09 0.24 0.00 0.03 0.00 0.00 1.34
## AU14_r_min 22 124 0.46 0.47 0.34 0.40 0.50 0.00 1.82
## AU15_r_min 23 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU17_r_min 24 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU20_r_min 25 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU23_r_min 26 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU25_r_min 27 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU26_r_min 28 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU45_r_min 29 124 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## lip_min 30 124 145.32 17.01 147.00 145.34 16.40 110.17 184.45
## eye_min 31 124 5.77 3.96 4.56 5.43 3.85 0.57 15.49
## amplitude_min 32 124 0.34 0.11 0.35 0.34 0.11 0.10 0.60
## stage_min 33 124 0.03 0.01 0.03 0.03 0.01 0.02 0.06
## apex_min 34 124 0.54 0.13 0.54 0.54 0.14 0.27 0.83
## offset_min 35 124 0.36 0.12 0.38 0.36 0.12 0.12 0.61
## onset_min 36 124 0.36 0.12 0.36 0.36 0.11 0.10 0.73
## frame_max 37 124 163.86 53.03 153.00 157.38 40.03 90.00 385.00
## timestamp_max 38 124 3.26 1.06 3.04 3.13 0.80 1.78 7.68
## gaze_angle_x_max 39 124 0.24 0.08 0.24 0.24 0.07 0.02 0.46
## gaze_angle_y_max 40 124 0.41 0.11 0.40 0.40 0.11 0.10 0.69
## pose_Rx_max 41 124 0.19 0.10 0.21 0.20 0.09 -0.10 0.36
## pose_Ry_max 42 124 -0.14 0.08 -0.14 -0.14 0.08 -0.38 0.04
## pose_Rz_max 43 124 0.03 0.08 0.03 0.02 0.06 -0.16 0.28
## AU01_r_max 44 124 0.84 0.72 0.72 0.71 0.42 0.13 5.00
## AU02_r_max 45 124 0.56 0.39 0.47 0.48 0.19 0.15 3.09
## AU04_r_max 46 124 0.46 0.71 0.00 0.31 0.01 0.00 2.83
## AU05_r_max 47 124 0.53 0.36 0.46 0.48 0.26 0.12 2.30
## AU06_r_max 48 124 1.50 0.77 1.50 1.50 0.72 0.00 3.53
## AU07_r_max 49 124 1.50 0.91 1.46 1.45 0.76 0.00 4.40
## AU09_r_max 50 124 0.32 0.19 0.26 0.29 0.13 0.10 0.98
## AU10_r_max 51 124 0.81 0.67 0.69 0.77 1.02 0.00 2.28
## AU12_r_max 52 124 2.66 0.72 2.70 2.65 0.74 0.79 4.70
## AU14_r_max 53 124 1.85 0.46 1.86 1.86 0.51 0.83 2.89
## AU15_r_max 54 124 0.42 0.23 0.36 0.38 0.17 0.13 1.27
## AU17_r_max 55 124 1.17 0.63 1.10 1.10 0.61 0.29 3.97
## AU20_r_max 56 124 0.42 0.23 0.36 0.39 0.20 0.10 1.46
## AU23_r_max 57 124 0.52 0.34 0.46 0.47 0.27 0.11 1.61
## AU25_r_max 58 124 1.32 0.76 1.13 1.26 0.73 0.32 3.30
## AU26_r_max 59 124 1.09 0.48 1.00 1.04 0.39 0.38 2.93
## AU45_r_max 60 124 1.93 1.34 1.98 1.85 1.93 0.22 5.00
## lip_max 61 124 189.14 20.68 191.42 188.83 22.46 144.05 239.26
## eye_max 62 124 12.98 2.26 12.92 12.95 2.01 7.83 19.03
## amplitude_max 63 124 0.63 0.14 0.64 0.63 0.15 0.33 0.96
## stage_max 64 124 0.07 0.03 0.06 0.06 0.02 0.03 0.20
## apex_max 65 124 0.63 0.14 0.64 0.63 0.15 0.33 0.96
## offset_max 66 124 0.52 0.13 0.53 0.52 0.13 0.26 0.82
## onset_max 67 124 0.52 0.12 0.53 0.53 0.13 0.26 0.81
## frame_mean 68 124 82.43 26.52 77.00 79.19 20.02 45.50 193.00
## timestamp_mean 69 124 1.63 0.53 1.52 1.56 0.40 0.89 3.84
## gaze_angle_x_mean 70 124 0.19 0.08 0.21 0.20 0.06 -0.04 0.41
## gaze_angle_y_mean 71 124 0.28 0.08 0.30 0.29 0.07 0.05 0.46
## pose_Rx_mean 72 124 0.15 0.10 0.17 0.16 0.10 -0.11 0.32
## pose_Ry_mean 73 124 -0.17 0.08 -0.17 -0.17 0.07 -0.41 0.04
## pose_Rz_mean 74 124 -0.01 0.07 0.00 -0.01 0.06 -0.18 0.16
## AU01_r_mean 75 124 0.15 0.16 0.11 0.12 0.06 0.03 1.06
## AU02_r_mean 76 124 0.06 0.05 0.05 0.05 0.02 0.02 0.37
## AU04_r_mean 77 124 0.24 0.52 0.00 0.10 0.00 0.00 2.40
## AU05_r_mean 78 124 0.05 0.03 0.04 0.04 0.02 0.01 0.24
## AU06_r_mean 79 124 0.80 0.53 0.71 0.77 0.53 0.00 2.48
## AU07_r_mean 80 124 0.78 0.68 0.66 0.68 0.58 0.00 3.02
## AU09_r_mean 81 124 0.04 0.02 0.03 0.04 0.02 0.01 0.15
## AU10_r_mean 82 124 0.36 0.36 0.25 0.31 0.37 0.00 1.46
## AU12_r_mean 83 124 1.73 0.56 1.84 1.74 0.55 0.31 3.25
## AU14_r_mean 84 124 1.31 0.44 1.36 1.33 0.44 0.31 2.20
## AU15_r_mean 85 124 0.07 0.04 0.06 0.07 0.02 0.03 0.36
## AU17_r_mean 86 124 0.35 0.19 0.31 0.33 0.19 0.08 1.02
## AU20_r_mean 87 124 0.06 0.03 0.05 0.05 0.02 0.02 0.18
## AU23_r_mean 88 124 0.07 0.05 0.06 0.06 0.04 0.02 0.29
## AU25_r_mean 89 124 0.55 0.42 0.42 0.50 0.35 0.07 1.86
## AU26_r_mean 90 124 0.29 0.13 0.28 0.27 0.12 0.09 0.78
## AU45_r_mean 91 124 0.20 0.14 0.16 0.18 0.13 0.04 0.79
## lip_mean 92 124 174.86 18.90 175.85 174.96 20.62 134.86 219.12
## eye_mean 93 124 10.71 2.21 10.52 10.60 1.68 6.06 16.57
## amplitude_mean 94 124 0.53 0.13 0.54 0.53 0.14 0.27 0.83
## stage_mean 95 124 0.04 0.01 0.04 0.04 0.01 0.02 0.09
## apex_mean 96 124 0.60 0.14 0.61 0.60 0.14 0.31 0.89
## offset_mean 97 124 0.42 0.12 0.43 0.42 0.13 0.18 0.72
## onset_mean 98 124 0.42 0.12 0.43 0.42 0.12 0.12 0.75
## frame_sd 99 124 47.45 15.31 44.31 45.58 11.56 26.12 111.28
## timestamp_sd 100 124 0.95 0.31 0.89 0.91 0.23 0.52 2.23
## gaze_angle_x_sd 101 124 0.02 0.01 0.02 0.02 0.01 0.00 0.09
## gaze_angle_y_sd 102 124 0.04 0.02 0.04 0.04 0.02 0.01 0.12
## pose_Rx_sd 103 124 0.02 0.01 0.02 0.02 0.01 0.01 0.08
## pose_Ry_sd 104 124 0.01 0.01 0.01 0.01 0.01 0.00 0.06
## pose_Rz_sd 105 124 0.02 0.02 0.01 0.02 0.01 0.00 0.11
## AU01_r_sd 106 124 0.23 0.25 0.17 0.18 0.10 0.04 1.78
## AU02_r_sd 107 124 0.12 0.11 0.10 0.10 0.04 0.03 0.80
## AU04_r_sd 108 124 0.08 0.12 0.00 0.06 0.00 0.00 0.47
## AU05_r_sd 109 124 0.11 0.08 0.09 0.10 0.05 0.03 0.53
## AU06_r_sd 110 124 0.53 0.29 0.52 0.52 0.25 0.00 1.21
## AU07_r_sd 111 124 0.39 0.25 0.37 0.38 0.24 0.00 1.15
## AU09_r_sd 112 124 0.08 0.05 0.06 0.07 0.03 0.02 0.25
## AU10_r_sd 113 124 0.28 0.25 0.22 0.25 0.33 0.00 0.97
## AU12_r_sd 114 124 0.92 0.30 0.91 0.91 0.29 0.24 2.05
## AU14_r_sd 115 124 0.36 0.14 0.36 0.35 0.15 0.12 0.82
## AU15_r_sd 116 124 0.11 0.07 0.09 0.10 0.04 0.03 0.46
## AU17_r_sd 117 124 0.36 0.22 0.33 0.33 0.22 0.07 1.11
## AU20_r_sd 118 124 0.10 0.06 0.09 0.09 0.04 0.02 0.34
## AU23_r_sd 119 124 0.13 0.09 0.11 0.12 0.08 0.03 0.51
## AU25_r_sd 120 124 0.47 0.33 0.37 0.44 0.29 0.07 1.42
## AU26_r_sd 121 124 0.32 0.15 0.29 0.30 0.13 0.09 0.79
## AU45_r_sd 122 124 0.42 0.31 0.38 0.39 0.35 0.05 1.44
## lip_sd 123 124 14.20 4.84 13.58 13.93 4.17 2.84 31.30
## eye_sd 124 124 1.54 0.91 1.45 1.47 0.98 0.19 5.10
## amplitude_sd 125 124 0.09 0.03 0.09 0.09 0.03 0.02 0.21
## stage_sd 126 124 0.01 0.01 0.01 0.01 0.00 0.00 0.05
## apex_sd 127 124 0.02 0.01 0.02 0.02 0.01 0.01 0.07
## offset_sd 128 124 0.05 0.02 0.04 0.04 0.02 0.00 0.13
## onset_sd 129 124 0.06 0.03 0.05 0.05 0.02 0.00 0.15
## range skew kurtosis se
## filename* 123.00 0.00 -1.23 3.23
## subject 466.00 -0.04 -1.52 13.57
## gender* 0.00 NaN NaN 0.00
## age 9.00 1.11 0.89 0.19
## smile_type* 0.00 NaN NaN 0.00
## frame_min 0.00 NaN NaN 0.00
## timestamp_min 0.00 NaN NaN 0.00
## gaze_angle_x_min 0.49 -0.59 1.22 0.01
## gaze_angle_y_min 0.48 -0.50 0.10 0.01
## pose_Rx_min 0.62 -0.88 0.85 0.01
## pose_Ry_min 0.45 0.01 0.26 0.01
## pose_Rz_min 0.35 -0.64 -0.01 0.01
## AU01_r_min 0.00 NaN NaN 0.00
## AU02_r_min 0.00 NaN NaN 0.00
## AU04_r_min 2.18 3.45 11.17 0.04
## AU05_r_min 0.00 NaN NaN 0.00
## AU06_r_min 1.21 4.77 22.63 0.02
## AU07_r_min 2.36 4.01 16.74 0.04
## AU09_r_min 0.00 NaN NaN 0.00
## AU10_r_min 0.24 8.50 74.37 0.00
## AU12_r_min 1.34 3.12 10.00 0.02
## AU14_r_min 1.82 0.85 -0.20 0.04
## AU15_r_min 0.00 NaN NaN 0.00
## AU17_r_min 0.00 NaN NaN 0.00
## AU20_r_min 0.00 NaN NaN 0.00
## AU23_r_min 0.00 NaN NaN 0.00
## AU25_r_min 0.00 NaN NaN 0.00
## AU26_r_min 0.00 NaN NaN 0.00
## AU45_r_min 0.00 NaN NaN 0.00
## lip_min 74.28 -0.02 -0.47 1.53
## eye_min 14.91 0.64 -0.71 0.36
## amplitude_min 0.50 -0.02 -0.47 0.01
## stage_min 0.04 1.06 0.99 0.00
## apex_min 0.56 -0.06 -0.59 0.01
## offset_min 0.50 -0.12 -0.69 0.01
## onset_min 0.63 0.10 -0.07 0.01
## frame_max 295.00 1.45 2.64 4.76
## timestamp_max 5.90 1.45 2.64 0.10
## gaze_angle_x_max 0.44 -0.23 0.42 0.01
## gaze_angle_y_max 0.59 0.16 -0.09 0.01
## pose_Rx_max 0.46 -0.67 -0.26 0.01
## pose_Ry_max 0.43 -0.02 0.21 0.01
## pose_Rz_max 0.44 0.74 1.44 0.01
## AU01_r_max 4.87 2.99 11.42 0.06
## AU02_r_max 2.94 3.26 14.86 0.04
## AU04_r_max 2.83 1.60 1.69 0.06
## AU05_r_max 2.18 2.20 6.49 0.03
## AU06_r_max 3.53 0.15 -0.34 0.07
## AU07_r_max 4.40 0.49 0.22 0.08
## AU09_r_max 0.88 1.39 1.53 0.02
## AU10_r_max 2.28 0.27 -1.25 0.06
## AU12_r_max 3.91 0.04 0.18 0.06
## AU14_r_max 2.06 -0.09 -0.57 0.04
## AU15_r_max 1.14 1.59 2.34 0.02
## AU17_r_max 3.68 1.38 2.96 0.06
## AU20_r_max 1.36 1.57 3.24 0.02
## AU23_r_max 1.50 1.33 1.49 0.03
## AU25_r_max 2.98 0.64 -0.68 0.07
## AU26_r_max 2.55 1.10 1.22 0.04
## AU45_r_max 4.78 0.30 -1.18 0.12
## lip_max 95.21 0.08 -0.39 1.86
## eye_max 11.19 0.16 -0.03 0.20
## amplitude_max 0.64 0.08 -0.39 0.01
## stage_max 0.16 1.70 2.75 0.00
## apex_max 0.64 0.08 -0.39 0.01
## offset_max 0.56 0.00 -0.57 0.01
## onset_max 0.55 -0.07 -0.63 0.01
## frame_mean 147.50 1.45 2.64 2.38
## timestamp_mean 2.95 1.45 2.64 0.05
## gaze_angle_x_mean 0.45 -0.36 0.85 0.01
## gaze_angle_y_mean 0.41 -0.51 -0.06 0.01
## pose_Rx_mean 0.43 -0.71 -0.37 0.01
## pose_Ry_mean 0.44 0.03 0.32 0.01
## pose_Rz_mean 0.34 -0.05 0.00 0.01
## AU01_r_mean 1.03 3.91 16.81 0.01
## AU02_r_mean 0.36 4.13 20.44 0.00
## AU04_r_mean 2.40 2.64 6.40 0.05
## AU05_r_mean 0.23 2.68 10.76 0.00
## AU06_r_mean 2.48 0.62 0.24 0.05
## AU07_r_mean 3.02 1.22 1.49 0.06
## AU09_r_mean 0.14 1.58 3.19 0.00
## AU10_r_mean 1.46 0.79 -0.28 0.03
## AU12_r_mean 2.94 -0.16 -0.28 0.05
## AU14_r_mean 1.89 -0.23 -0.76 0.04
## AU15_r_mean 0.34 3.14 15.26 0.00
## AU17_r_mean 0.94 0.89 0.54 0.02
## AU20_r_mean 0.16 1.79 3.53 0.00
## AU23_r_mean 0.27 1.72 3.16 0.00
## AU25_r_mean 1.79 1.02 0.30 0.04
## AU26_r_mean 0.69 1.23 1.71 0.01
## AU45_r_mean 0.75 1.33 2.54 0.01
## lip_mean 84.25 -0.05 -0.58 1.70
## eye_mean 10.51 0.46 0.11 0.20
## amplitude_mean 0.56 -0.05 -0.58 0.01
## stage_mean 0.06 1.25 1.19 0.00
## apex_mean 0.59 0.05 -0.53 0.01
## offset_mean 0.54 -0.05 -0.67 0.01
## onset_mean 0.62 -0.06 -0.27 0.01
## frame_sd 85.16 1.45 2.64 1.37
## timestamp_sd 1.70 1.45 2.64 0.03
## gaze_angle_x_sd 0.09 2.87 11.79 0.00
## gaze_angle_y_sd 0.11 1.18 2.12 0.00
## pose_Rx_sd 0.08 1.78 3.79 0.00
## pose_Ry_sd 0.06 2.67 9.23 0.00
## pose_Rz_sd 0.11 2.21 5.01 0.00
## AU01_r_sd 1.74 3.75 16.21 0.02
## AU02_r_sd 0.77 3.78 17.96 0.01
## AU04_r_sd 0.47 1.29 0.57 0.01
## AU05_r_sd 0.51 2.32 7.43 0.01
## AU06_r_sd 1.21 0.19 -0.39 0.03
## AU07_r_sd 1.15 0.40 -0.09 0.02
## AU09_r_sd 0.23 1.33 1.27 0.00
## AU10_r_sd 0.97 0.60 -0.57 0.02
## AU12_r_sd 1.81 0.59 1.42 0.03
## AU14_r_sd 0.70 0.46 -0.12 0.01
## AU15_r_sd 0.43 2.17 6.17 0.01
## AU17_r_sd 1.04 1.11 1.27 0.02
## AU20_r_sd 0.31 1.74 3.45 0.01
## AU23_r_sd 0.49 1.54 2.47 0.01
## AU25_r_sd 1.35 0.81 -0.33 0.03
## AU26_r_sd 0.70 1.00 0.75 0.01
## AU45_r_sd 1.39 0.76 0.05 0.03
## lip_sd 28.46 0.68 1.23 0.43
## eye_sd 4.91 0.92 1.44 0.08
## amplitude_sd 0.19 0.68 1.23 0.00
## stage_sd 0.04 1.89 3.95 0.00
## apex_sd 0.07 2.06 7.82 0.00
## offset_sd 0.13 0.89 0.81 0.00
## onset_sd 0.15 0.79 0.77 0.00
## ------------------------------------------------------------
## smile_type: spontaneous
## gender: male
## vars n mean sd median trimmed mad min max
## filename* 1 122 61.50 35.36 61.50 61.50 45.22 1.00 122.00
## subject 2 122 290.67 143.53 272.00 289.95 197.93 54.00 525.00
## gender* 3 122 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## age 4 122 10.53 2.14 10.00 10.24 1.48 8.00 17.00
## smile_type* 5 122 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## frame_min 6 122 1.00 0.00 1.00 1.00 0.00 1.00 1.00
## timestamp_min 7 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## gaze_angle_x_min 8 122 0.16 0.09 0.17 0.16 0.09 -0.13 0.39
## gaze_angle_y_min 9 122 0.20 0.09 0.20 0.20 0.08 -0.13 0.38
## pose_Rx_min 10 122 0.08 0.11 0.09 0.09 0.11 -0.22 0.28
## pose_Ry_min 11 122 -0.21 0.10 -0.21 -0.21 0.09 -0.46 0.02
## pose_Rz_min 12 122 -0.05 0.08 -0.04 -0.04 0.07 -0.38 0.13
## AU01_r_min 13 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU02_r_min 14 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU04_r_min 15 122 0.14 0.38 0.00 0.03 0.00 0.00 1.92
## AU05_r_min 16 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU06_r_min 17 122 0.08 0.26 0.00 0.01 0.00 0.00 1.82
## AU07_r_min 18 122 0.21 0.49 0.00 0.08 0.00 0.00 2.50
## AU09_r_min 19 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU10_r_min 20 122 0.04 0.14 0.00 0.00 0.00 0.00 0.81
## AU12_r_min 21 122 0.17 0.33 0.00 0.10 0.00 0.00 1.95
## AU14_r_min 22 122 0.48 0.42 0.49 0.44 0.53 0.00 1.85
## AU15_r_min 23 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU17_r_min 24 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU20_r_min 25 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU23_r_min 26 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU25_r_min 27 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU26_r_min 28 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## AU45_r_min 29 122 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## lip_min 30 122 146.44 16.86 146.36 145.97 18.16 109.27 185.49
## eye_min 31 122 5.63 3.52 4.68 5.36 3.39 0.35 15.45
## amplitude_min 32 122 0.34 0.11 0.34 0.34 0.12 0.09 0.60
## stage_min 33 122 0.03 0.01 0.03 0.03 0.01 0.02 0.06
## apex_min 34 122 0.52 0.13 0.53 0.51 0.14 0.29 0.90
## offset_min 35 122 0.38 0.12 0.37 0.37 0.12 0.09 0.63
## onset_min 36 122 0.36 0.11 0.37 0.36 0.12 0.15 0.74
## frame_max 37 122 251.30 110.38 226.50 241.56 101.56 55.00 705.00
## timestamp_max 38 122 5.01 2.21 4.51 4.81 2.03 1.08 14.08
## gaze_angle_x_max 39 122 0.25 0.08 0.25 0.25 0.08 0.02 0.47
## gaze_angle_y_max 40 122 0.39 0.12 0.40 0.40 0.13 0.07 0.62
## pose_Rx_max 41 122 0.21 0.10 0.21 0.21 0.11 -0.06 0.45
## pose_Ry_max 42 122 -0.14 0.10 -0.14 -0.14 0.10 -0.40 0.11
## pose_Rz_max 43 122 0.04 0.07 0.04 0.04 0.07 -0.11 0.26
## AU01_r_max 44 122 0.76 0.63 0.56 0.63 0.34 0.23 4.52
## AU02_r_max 45 122 0.60 0.41 0.47 0.52 0.21 0.16 2.04
## AU04_r_max 46 122 0.61 0.88 0.21 0.43 0.31 0.00 3.56
## AU05_r_max 47 122 0.51 0.26 0.46 0.48 0.22 0.15 1.85
## AU06_r_max 48 122 1.69 0.79 1.63 1.68 0.73 0.00 3.74
## AU07_r_max 49 122 1.84 1.03 1.66 1.78 0.90 0.00 5.00
## AU09_r_max 50 122 0.40 0.25 0.34 0.35 0.18 0.12 1.55
## AU10_r_max 51 122 1.22 0.70 1.31 1.23 0.74 0.00 2.87
## AU12_r_max 52 122 2.70 0.60 2.71 2.72 0.57 1.23 4.05
## AU14_r_max 53 122 1.79 0.50 1.81 1.80 0.43 0.03 3.43
## AU15_r_max 54 122 0.49 0.29 0.41 0.45 0.16 0.19 2.76
## AU17_r_max 55 122 1.22 0.55 1.09 1.17 0.48 0.38 2.82
## AU20_r_max 56 122 0.49 0.21 0.44 0.47 0.20 0.12 1.14
## AU23_r_max 57 122 0.61 0.38 0.53 0.56 0.32 0.10 1.91
## AU25_r_max 58 122 1.25 0.69 1.04 1.16 0.63 0.38 3.46
## AU26_r_max 59 122 1.20 0.68 1.02 1.11 0.41 0.29 4.73
## AU45_r_max 60 122 1.74 1.11 1.55 1.68 1.48 0.25 3.96
## lip_max 61 122 185.98 20.29 187.39 185.88 25.78 146.48 244.90
## eye_max 62 122 12.51 2.43 12.45 12.43 2.25 7.25 22.05
## amplitude_max 63 122 0.61 0.14 0.62 0.61 0.17 0.34 1.00
## stage_max 64 122 0.07 0.03 0.06 0.06 0.02 0.03 0.21
## apex_max 65 122 0.61 0.14 0.62 0.61 0.17 0.34 1.00
## offset_max 66 122 0.51 0.13 0.52 0.51 0.14 0.29 0.90
## onset_max 67 122 0.51 0.13 0.52 0.51 0.14 0.28 0.89
## frame_mean 68 122 126.15 55.19 113.75 121.28 50.78 28.00 353.00
## timestamp_mean 69 122 2.50 1.10 2.26 2.41 1.02 0.54 7.04
## gaze_angle_x_mean 70 122 0.21 0.08 0.21 0.21 0.08 -0.01 0.43
## gaze_angle_y_mean 71 122 0.27 0.08 0.28 0.27 0.08 0.00 0.45
## pose_Rx_mean 72 122 0.15 0.10 0.17 0.16 0.09 -0.11 0.34
## pose_Ry_mean 73 122 -0.18 0.09 -0.18 -0.18 0.09 -0.43 0.05
## pose_Rz_mean 74 122 0.00 0.07 -0.01 0.00 0.07 -0.18 0.19
## AU01_r_mean 75 122 0.12 0.11 0.09 0.10 0.05 0.04 0.95
## AU02_r_mean 76 122 0.06 0.05 0.05 0.05 0.02 0.02 0.29
## AU04_r_mean 77 122 0.32 0.60 0.01 0.17 0.01 0.00 2.70
## AU05_r_mean 78 122 0.04 0.02 0.04 0.04 0.01 0.01 0.12
## AU06_r_mean 79 122 0.93 0.63 0.77 0.88 0.52 0.00 2.95
## AU07_r_mean 80 122 0.99 0.82 0.81 0.89 0.68 0.00 3.63
## AU09_r_mean 81 122 0.05 0.04 0.04 0.04 0.02 0.01 0.30
## AU10_r_mean 82 122 0.61 0.50 0.54 0.57 0.55 0.00 1.84
## AU12_r_mean 83 122 1.79 0.58 1.78 1.78 0.54 0.40 3.38
## AU14_r_mean 84 122 1.23 0.46 1.23 1.24 0.39 0.00 2.81
## AU15_r_mean 85 122 0.08 0.04 0.07 0.07 0.02 0.03 0.30
## AU17_r_mean 86 122 0.32 0.15 0.27 0.30 0.12 0.12 0.87
## AU20_r_mean 87 122 0.06 0.03 0.06 0.06 0.02 0.02 0.16
## AU23_r_mean 88 122 0.08 0.05 0.06 0.07 0.04 0.01 0.26
## AU25_r_mean 89 122 0.43 0.28 0.35 0.40 0.24 0.12 1.70
## AU26_r_mean 90 122 0.32 0.19 0.28 0.29 0.13 0.10 1.41
## AU45_r_mean 91 122 0.16 0.10 0.13 0.15 0.08 0.05 0.48
## lip_mean 92 122 172.02 18.90 173.39 171.86 21.21 138.57 229.63
## eye_mean 93 122 10.36 2.25 9.99 10.21 2.29 5.80 17.31
## amplitude_mean 94 122 0.51 0.13 0.52 0.51 0.14 0.29 0.90
## stage_mean 95 122 0.04 0.01 0.04 0.04 0.01 0.02 0.08
## apex_mean 96 122 0.57 0.13 0.58 0.57 0.15 0.32 0.96
## offset_mean 97 122 0.44 0.12 0.45 0.44 0.13 0.21 0.73
## onset_mean 98 122 0.43 0.12 0.44 0.43 0.13 0.21 0.76
## frame_sd 99 122 72.69 31.86 65.53 69.88 29.32 16.02 203.66
## timestamp_sd 100 122 1.45 0.64 1.31 1.40 0.59 0.32 4.07
## gaze_angle_x_sd 101 122 0.02 0.01 0.02 0.02 0.01 0.00 0.09
## gaze_angle_y_sd 102 122 0.04 0.02 0.04 0.04 0.02 0.01 0.10
## pose_Rx_sd 103 122 0.03 0.02 0.02 0.03 0.01 0.01 0.11
## pose_Ry_sd 104 122 0.02 0.01 0.01 0.02 0.01 0.00 0.07
## pose_Rz_sd 105 122 0.02 0.02 0.02 0.02 0.01 0.00 0.11
## AU01_r_sd 106 122 0.19 0.19 0.14 0.15 0.09 0.05 1.60
## AU02_r_sd 107 122 0.13 0.10 0.10 0.10 0.05 0.04 0.54
## AU04_r_sd 108 122 0.12 0.17 0.03 0.08 0.04 0.00 0.67
## AU05_r_sd 109 122 0.10 0.05 0.09 0.09 0.04 0.03 0.34
## AU06_r_sd 110 122 0.50 0.21 0.49 0.50 0.20 0.00 0.99
## AU07_r_sd 111 122 0.41 0.21 0.40 0.40 0.17 0.00 1.03
## AU09_r_sd 112 122 0.09 0.07 0.07 0.08 0.04 0.02 0.54
## AU10_r_sd 113 122 0.35 0.19 0.38 0.35 0.17 0.00 0.74
## AU12_r_sd 114 122 0.76 0.19 0.75 0.75 0.20 0.32 1.34
## AU14_r_sd 115 122 0.30 0.13 0.26 0.29 0.10 0.01 0.83
## AU15_r_sd 116 122 0.12 0.07 0.10 0.11 0.04 0.05 0.64
## AU17_r_sd 117 122 0.32 0.15 0.29 0.31 0.14 0.10 0.95
## AU20_r_sd 118 122 0.11 0.05 0.10 0.11 0.05 0.03 0.30
## AU23_r_sd 119 122 0.14 0.09 0.12 0.13 0.08 0.03 0.47
## AU25_r_sd 120 122 0.41 0.27 0.31 0.37 0.21 0.11 1.43
## AU26_r_sd 121 122 0.33 0.20 0.28 0.31 0.13 0.09 1.46
## AU45_r_sd 122 122 0.33 0.21 0.28 0.30 0.21 0.06 0.92
## lip_sd 123 122 11.13 3.44 11.07 10.98 3.58 4.95 19.10
## eye_sd 124 122 1.33 0.67 1.31 1.29 0.75 0.23 3.19
## amplitude_sd 125 122 0.07 0.02 0.07 0.07 0.02 0.03 0.13
## stage_sd 126 122 0.01 0.00 0.01 0.01 0.00 0.00 0.02
## apex_sd 127 122 0.03 0.01 0.03 0.02 0.01 0.01 0.05
## offset_sd 128 122 0.04 0.02 0.04 0.04 0.02 0.00 0.13
## onset_sd 129 122 0.04 0.02 0.04 0.04 0.02 0.01 0.11
## range skew kurtosis se
## filename* 121.00 0.00 -1.23 3.20
## subject 471.00 0.14 -1.37 12.99
## gender* 0.00 NaN NaN 0.00
## age 9.00 1.19 0.82 0.19
## smile_type* 0.00 NaN NaN 0.00
## frame_min 0.00 NaN NaN 0.00
## timestamp_min 0.00 NaN NaN 0.00
## gaze_angle_x_min 0.52 -0.52 1.19 0.01
## gaze_angle_y_min 0.51 -0.54 0.85 0.01
## pose_Rx_min 0.51 -0.50 -0.34 0.01
## pose_Ry_min 0.48 -0.14 -0.28 0.01
## pose_Rz_min 0.50 -0.56 1.09 0.01
## AU01_r_min 0.00 NaN NaN 0.00
## AU02_r_min 0.00 NaN NaN 0.00
## AU04_r_min 1.92 3.16 9.76 0.03
## AU05_r_min 0.00 NaN NaN 0.00
## AU06_r_min 1.82 4.31 20.80 0.02
## AU07_r_min 2.50 3.00 9.06 0.04
## AU09_r_min 0.00 NaN NaN 0.00
## AU10_r_min 0.81 4.21 17.09 0.01
## AU12_r_min 1.95 2.87 9.64 0.03
## AU14_r_min 1.85 0.61 -0.16 0.04
## AU15_r_min 0.00 NaN NaN 0.00
## AU17_r_min 0.00 NaN NaN 0.00
## AU20_r_min 0.00 NaN NaN 0.00
## AU23_r_min 0.00 NaN NaN 0.00
## AU25_r_min 0.00 NaN NaN 0.00
## AU26_r_min 0.00 NaN NaN 0.00
## AU45_r_min 0.00 NaN NaN 0.00
## lip_min 76.23 0.20 -0.61 1.53
## eye_min 15.10 0.67 -0.41 0.32
## amplitude_min 0.51 0.20 -0.61 0.01
## stage_min 0.04 0.67 -0.47 0.00
## apex_min 0.61 0.15 -0.52 0.01
## offset_min 0.54 0.18 -0.67 0.01
## onset_min 0.59 0.29 -0.14 0.01
## frame_max 650.00 1.13 2.16 9.99
## timestamp_max 13.00 1.13 2.16 0.20
## gaze_angle_x_max 0.44 0.01 0.21 0.01
## gaze_angle_y_max 0.55 -0.23 -0.40 0.01
## pose_Rx_max 0.51 -0.40 -0.14 0.01
## pose_Ry_max 0.51 0.00 -0.23 0.01
## pose_Rz_max 0.37 0.29 -0.16 0.01
## AU01_r_max 4.29 2.94 11.44 0.06
## AU02_r_max 1.88 1.90 3.26 0.04
## AU04_r_max 3.56 1.57 1.59 0.08
## AU05_r_max 1.70 1.76 5.44 0.02
## AU06_r_max 3.74 0.19 0.00 0.07
## AU07_r_max 5.00 0.68 0.46 0.09
## AU09_r_max 1.43 2.23 6.41 0.02
## AU10_r_max 2.87 -0.12 -0.75 0.06
## AU12_r_max 2.82 -0.21 -0.19 0.05
## AU14_r_max 3.40 -0.24 1.32 0.05
## AU15_r_max 2.57 4.07 26.86 0.03
## AU17_r_max 2.44 0.82 0.06 0.05
## AU20_r_max 1.02 0.83 0.18 0.02
## AU23_r_max 1.81 1.23 1.36 0.03
## AU25_r_max 3.08 1.02 0.49 0.06
## AU26_r_max 4.44 2.10 6.50 0.06
## AU45_r_max 3.71 0.27 -1.34 0.10
## lip_max 98.42 0.09 -0.64 1.84
## eye_max 14.80 0.71 1.84 0.22
## amplitude_max 0.66 0.09 -0.64 0.01
## stage_max 0.18 1.82 5.60 0.00
## apex_max 0.66 0.09 -0.64 0.01
## offset_max 0.61 0.12 -0.57 0.01
## onset_max 0.61 0.13 -0.58 0.01
## frame_mean 325.00 1.13 2.16 5.00
## timestamp_mean 6.50 1.13 2.16 0.10
## gaze_angle_x_mean 0.44 -0.01 0.33 0.01
## gaze_angle_y_mean 0.46 -0.53 0.56 0.01
## pose_Rx_mean 0.45 -0.55 -0.19 0.01
## pose_Ry_mean 0.48 -0.09 -0.26 0.01
## pose_Rz_mean 0.36 0.10 -0.35 0.01
## AU01_r_mean 0.92 4.11 22.90 0.01
## AU02_r_mean 0.27 2.84 8.41 0.00
## AU04_r_mean 2.70 2.20 4.33 0.05
## AU05_r_mean 0.11 1.79 4.08 0.00
## AU06_r_mean 2.95 0.84 0.42 0.06
## AU07_r_mean 3.63 1.21 1.39 0.07
## AU09_r_mean 0.29 3.57 16.95 0.00
## AU10_r_mean 1.84 0.58 -0.68 0.04
## AU12_r_mean 2.98 0.08 -0.06 0.05
## AU14_r_mean 2.80 -0.10 0.77 0.04
## AU15_r_mean 0.26 2.41 8.85 0.00
## AU17_r_mean 0.74 1.33 1.70 0.01
## AU20_r_mean 0.14 0.98 0.80 0.00
## AU23_r_mean 0.25 1.26 1.73 0.00
## AU25_r_mean 1.58 1.38 2.45 0.03
## AU26_r_mean 1.32 2.69 10.97 0.02
## AU45_r_mean 0.44 1.36 1.75 0.01
## lip_mean 91.06 0.14 -0.53 1.71
## eye_mean 11.51 0.75 0.54 0.20
## amplitude_mean 0.61 0.14 -0.53 0.01
## stage_mean 0.05 0.80 -0.27 0.00
## apex_mean 0.64 0.13 -0.59 0.01
## offset_mean 0.53 0.04 -0.81 0.01
## onset_mean 0.55 0.06 -0.73 0.01
## frame_sd 187.64 1.13 2.16 2.88
## timestamp_sd 3.75 1.13 2.16 0.06
## gaze_angle_x_sd 0.08 2.50 8.79 0.00
## gaze_angle_y_sd 0.09 1.06 1.35 0.00
## pose_Rx_sd 0.11 1.44 2.43 0.00
## pose_Ry_sd 0.06 1.71 3.43 0.00
## pose_Rz_sd 0.10 2.03 5.09 0.00
## AU01_r_sd 1.55 4.15 23.39 0.02
## AU02_r_sd 0.50 2.46 6.03 0.01
## AU04_r_sd 0.67 1.57 1.72 0.02
## AU05_r_sd 0.31 1.90 4.90 0.00
## AU06_r_sd 0.99 -0.29 0.09 0.02
## AU07_r_sd 1.03 0.29 0.21 0.02
## AU09_r_sd 0.52 3.18 14.01 0.01
## AU10_r_sd 0.74 -0.31 -0.78 0.02
## AU12_r_sd 1.02 0.29 -0.19 0.02
## AU14_r_sd 0.82 0.91 1.49 0.01
## AU15_r_sd 0.60 3.67 22.46 0.01
## AU17_r_sd 0.84 1.04 1.42 0.01
## AU20_r_sd 0.27 0.96 0.67 0.00
## AU23_r_sd 0.45 1.11 0.92 0.01
## AU25_r_sd 1.33 1.18 0.94 0.02
## AU26_r_sd 1.38 2.34 8.78 0.02
## AU45_r_sd 0.85 0.86 -0.11 0.02
## lip_sd 14.15 0.35 -0.45 0.31
## eye_sd 2.96 0.54 -0.21 0.06
## amplitude_sd 0.09 0.35 -0.45 0.00
## stage_sd 0.02 1.22 0.76 0.00
## apex_sd 0.05 0.60 0.11 0.00
## offset_sd 0.13 1.33 2.69 0.00
## onset_sd 0.11 0.79 0.28 0.00
# citation("psych")
The main conclusions of the descriptive statistics will be given in the thesis report. The ggplot package is used to create distribution visualization of the distribution of the features.
# loading packages
library(ggplot2)
library(ggpubr)
# age & gender distribution of the video's
ggplot(UvA_sum, aes(y = age, x = smile_type, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "Age") +
scale_x_discrete(name = "Smile Type")
fig1 <- ggplot(UvA_sum, aes(y = age, x = smile_type, color = gender)) +
geom_boxplot() +
scale_y_continuous(name = "Age") +
scale_x_discrete(name = "Smile Type")
ggplot(data = UvA_sum, aes(x = age, fill = smile_type)) +
geom_bar(stat = "count") +
scale_x_continuous(
name = "Age",
breaks = c(8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
) +
scale_y_continuous(name = "Smile Type") +
labs() +
theme(
legend.position = "bottom", text = element_text(size = 10),
axis.text = element_text(size = 10)
)
ggplot(data = UvA_sum, aes(x = age, fill = gender)) +
geom_bar(stat = "count") +
scale_x_continuous(
name = "Age",
breaks = c(8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
) +
scale_y_continuous(name = "Gender") +
labs() +
theme(
legend.position = "bottom", text = element_text(size = 10),
axis.text = element_text(size = 10)
)
fig2 <- ggplot(data = UvA_sum, aes(x = age, fill = smile_type)) +
geom_bar(stat = "count") +
scale_x_continuous(
name = "Age",
breaks = c(8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
) +
scale_y_continuous(name = "Smile Type") +
facet_grid(smile_type ~ gender) +
labs() +
theme(
legend.position = "bottom", text = element_text(size = 10),
axis.text = element_text(size = 10)
)
# use the ggpubr package to combine multiple ggplot visualizations in one plot
figure <- ggarrange(fig1, fig2,
# labels = c("1", "2"),
ncol = 2, nrow = 1
)
figure
# citation("ggpubr")
Check the distributions and any possible outliers, using the mean() and sd().
# distribution per feature
# age
ggplot(UvA_sum, aes(x = age)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = age, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
# frame time
ggplot(UvA_sum, aes(x = frame_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = frame_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4) +
scale_x_continuous(name = "Frame Mean") +
scale_y_continuous(name = "Smile Type Count") +
theme(
legend.position = "bottom", text = element_text(size = 10),
axis.text = element_text(size = 10)
)
# gaze_angle_x
ggplot(UvA_sum, aes(x = gaze_angle_x_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = gaze_angle_x_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(
UvA_sum,
aes(x = smile_type, y = gaze_angle_x_mean, color = smile_type)
) +
geom_boxplot() +
scale_y_continuous(name = "Gaze x") +
scale_x_discrete(name = "Smile Type")
# gaze_angle_y
ggplot(UvA_sum, aes(x = gaze_angle_y_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = gaze_angle_y_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(
UvA_sum,
aes(x = smile_type, y = gaze_angle_y_mean, color = smile_type)
) +
geom_boxplot() +
scale_y_continuous(name = "Gaze y") +
scale_x_discrete(name = "Smile Type")
# pose_Rx
ggplot(UvA_sum, aes(x = pose_Rx_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = pose_Rx_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = smile_type, y = pose_Rx_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "Pose Rx") +
scale_x_discrete(name = "Smile Type")
# pose_Ry
ggplot(UvA_sum, aes(x = pose_Ry_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = pose_Ry_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = smile_type, y = pose_Ry_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "Pose Ry") +
scale_x_discrete(name = "Smile Type")
# pose_Rz
ggplot(UvA_sum, aes(x = pose_Rz_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = pose_Rz_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = smile_type, y = pose_Rz_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "Pose Rz") +
scale_x_discrete(name = "Smile Type")
# AU01
ggplot(UvA_sum, aes(x = AU01_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU01_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU01_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU01_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU01_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU01") +
scale_x_discrete(name = "Smile Type")
# AU02
ggplot(UvA_sum, aes(x = AU02_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU02_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU02_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU02_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU02_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU02") +
scale_x_discrete(name = "Smile Type")
# AU04
ggplot(UvA_sum, aes(x = AU04_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU04_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
# AU04 does not seem to have an impact, as most values show zero value
# AU05
ggplot(UvA_sum, aes(x = AU05_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU05_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU05_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU05_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU05_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU05") +
scale_x_discrete(name = "Smile Type")
# AU06
ggplot(UvA_sum, aes(x = AU06_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU06_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU06_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU06_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU06_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU06") +
scale_x_discrete(name = "Smile Type")
# AU07
ggplot(UvA_sum, aes(x = AU07_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU07_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU07_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU07_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU07_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU07") +
scale_x_discrete(name = "Smile Type")
# AU09
ggplot(UvA_sum, aes(x = AU09_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU09_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU09_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU09_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU02_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU01") +
scale_x_discrete(name = "Smile Type")
# AU10
ggplot(UvA_sum, aes(x = AU10_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU10_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU10_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU10_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU10_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU10") +
scale_x_discrete(name = "Smile Type")
# AU12
ggplot(UvA_sum, aes(x = AU12_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
fig5 <- ggplot(UvA_sum, aes(x = AU12_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU12_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
fig4 <- ggplot(UvA_sum, aes(x = AU12_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
fig3 <- ggplot(
UvA_sum,
aes(x = smile_type, y = AU12_r_mean, color = smile_type)
) +
geom_boxplot() +
scale_y_continuous(name = "AU12") +
scale_x_discrete(name = "Smile Type")
figure1 <- ggarrange(fig3, fig4, fig5,
# labels = c("1", "2"),
ncol = 2, nrow = 2
)
figure1
# AU14
ggplot(UvA_sum, aes(x = AU14_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU14_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU14_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU14_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU14_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU14") +
scale_x_discrete(name = "Smile Type")
# AU15
ggplot(UvA_sum, aes(x = AU15_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU15_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU15_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU15_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU15_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU15") +
scale_x_discrete(name = "Smile Type")
# AU17
ggplot(UvA_sum, aes(x = AU17_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU17_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU17_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU17_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU17_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU17") +
scale_x_discrete(name = "Smile Type")
# AU20
ggplot(UvA_sum, aes(x = AU20_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU20_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU20_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU20_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU20_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU20") +
scale_x_discrete(name = "Smile Type")
# AU23
ggplot(UvA_sum, aes(x = AU23_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU23_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU23_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU23_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU23_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU23") +
scale_x_discrete(name = "Smile Type")
# AU25
ggplot(UvA_sum, aes(x = AU25_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
fig8 <- ggplot(UvA_sum, aes(x = AU25_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU25_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
fig7 <- ggplot(UvA_sum, aes(x = AU25_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
fig6 <- ggplot(
UvA_sum,
aes(x = smile_type, y = AU25_r_mean, color = smile_type)
) +
geom_boxplot() +
scale_y_continuous(name = "AU25") +
scale_x_discrete(name = "Smile Type")
figure2 <- ggarrange(fig6, fig7, fig8,
# labels = c("1", "2"),
ncol = 2, nrow = 2
)
figure2
# AU26
ggplot(UvA_sum, aes(x = AU26_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU26_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU26_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU26_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU26_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU26") +
scale_x_discrete(name = "Smile Type")
# AU45
ggplot(UvA_sum, aes(x = AU45_r_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU45_r_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = AU45_r_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = AU45_r_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = AU45_r_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU45") +
scale_x_discrete(name = "Smile Type")
# lip
ggplot(UvA_sum, aes(x = lip_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = lip_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = lip_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = lip_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = lip_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "Lip") +
scale_x_discrete(name = "Smile Type")
# eye
ggplot(UvA_sum, aes(x = eye_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = eye_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = eye_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = eye_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = eye_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "eye") +
scale_x_discrete(name = "Smile Type")
# amplitude for onset, apex and offset
ggplot(UvA_sum, aes(x = amplitude_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = amplitude_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = amplitude_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = amplitude_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = amplitude_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "amlitude") +
scale_x_discrete(name = "Smile Type")
# apex
ggplot(UvA_sum, aes(x = apex_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = apex_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = apex_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = apex_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = apex_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU01") +
scale_x_discrete(name = "Smile Type")
# onset
ggplot(UvA_sum, aes(x = onset_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = onset_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = onset_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = onset_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = onset_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "AU01") +
scale_x_discrete(name = "Smile Type")
# offset
ggplot(UvA_sum, aes(x = offset_sd)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = offset_sd, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = offset_mean)) +
geom_histogram(fill = "white", colour = "black") +
facet_grid(smile_type ~ ., scales = "free")
ggplot(UvA_sum, aes(x = offset_mean, fill = smile_type)) +
geom_histogram(position = "identity", alpha = 0.4)
ggplot(UvA_sum, aes(x = smile_type, y = offset_mean, color = smile_type)) +
geom_boxplot() +
scale_y_continuous(name = "offset") +
scale_x_discrete(name = "Smile Type")
Some other check plot types to be maybe used later on
# apex vs AU's 6 and 12 for happiness
ggplot(data = UvA_sum, aes(x = AU12_r_mean, y = AU06_r_mean)) +
geom_point() +
geom_smooth() +
scale_x_continuous(
name = "AU12 ",
) +
scale_y_continuous(name = "AU06") +
labs() +
theme(
legend.position = "bottom", text = element_text(size = 10),
axis.text = element_text(size = 10)
)
## `geom_smooth()` using method = 'loess' and formula 'y ~ x'
# correlation between eye and lip
ggplot(data = UvA_sum, aes(y = eye_mean, x = lip_mean)) +
geom_point() +
geom_smooth()
## `geom_smooth()` using method = 'loess' and formula 'y ~ x'
# correlation between onset and apex
ggplot(data = UvA_sum, aes(y = onset_mean, x = apex_mean)) +
geom_point() +
geom_smooth()
## `geom_smooth()` using method = 'loess' and formula 'y ~ x'
The caret package is used to perform the training, testing and evaluation as well the splitting the data. Further explanation of the options and parameter settings of the caret package can be found in the thesis, via the citation link, or ? R help function.
# set.seed for modeling to 1973 for all models and predictions
set.seed(1973)
# loading packages
library(caret)
library(ggplot2)
library(dplyr)
library(tree)
library(rpart)
library(rpart.plot)
library(rattle)
# remove the filename (or ID) from the modelset to avoid overfilling
UvA_modelset$filename <- NULL
UvA_modelset$smile_type <- as.factor(UvA_modelset$smile_type)
UvA_modelset$gender <- as.factor(UvA_modelset$gender)
# relevel to spontaneous smile
UvA_modelset$smile_type <- relevel(UvA_modelset$smile_type, ref = "spontaneous")
# Split into training and test
set.seed(1973)
trn_index <- createDataPartition(UvA_modelset$smile_type, p = 0.7, list = FALSE)
trn_smile <- UvA_modelset[trn_index, ]
tst_smile <- UvA_modelset[-trn_index, ]
# Split the test set into boys and girls for detecting differences
set.seed(1973)
tst_smile_girls <- tst_smile %>%
filter(gender == "female")
tst_smile_boys <- tst_smile %>%
filter(gender == "male ")
# check the balance in the dataset for the independent variable
table(UvA_modelset$smile_type)
##
## spontaneous deliberate
## 235 240
# citation("caret")
# citation("tree")
# citation("rpart")
# citation("rpart.plot")
# citation("ggplot2")
# citation("rattle")
For the decision trees two packages are explored. For convenience of the project the choice has been made to work with the rpart package over the tree package. More information about these two packages can be found in the citation link or ? R help function. The trained models are divided into eight categories. Multiple models per category are explored. The explanation on the categories can be found in the thesis. To train the models the train() function is used. The models are stored as variable. The parameter settings are explained in the thesis. The models use 10 fold cross-validation. To check the density a visualization is added to the complete model. On this first complete model, also the pre-processing parameter is tested. This is done to see if scaling and centering the the dependent features improves the model. This is not the case for the complete decision tree including all features. The parameter is kept at default for that reason. To visualize the trained decision trees the rattle package is used. The package provides a nicer looking tree. The predict() function is used to create the predictions based on the test set, and stored as variable. For model evaluation the confusionMatrix() function is used and printed.
# check the balance for the baseline model
baseline_model <- table(trn_smile$smile_type)
baseline_model
##
## spontaneous deliberate
## 165 168
# model 0: complete model
# set the seed
set.seed(1973)
# train the model using train and rpart, store the model
smile__tree_model_0 <- train(smile_type ~ .,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
# check the outcome of the model
smile__tree_model_0$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5370544 0.07672893 0.07505544 0.1510397
## 2 0.05252525 0.5369652 0.07274720 0.07414327 0.1465495
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
# check and visualize the variable importance
varImp_0 <- varImp(smile__tree_model_0)
varImp_0
## rpart variable importance
##
## only 20 most important variables shown (out of 32)
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 65.43
## AU25_r_mean 63.30
## apex_mean 59.99
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## amplitude_mean 43.03
## lip_mean 43.03
## AU09_r_mean 40.62
## stage_mean 24.72
## AU14_r_mean 21.58
## AU20_r_mean 15.81
## AU15_r_mean 0.00
## age 0.00
## onset_mean 0.00
## AU23_r_mean 0.00
## AU12_r_mean 0.00
## pose_Ry_mean 0.00
## `gendermale ` 0.00
## AU04_r_mean 0.00
par(mfrow = c(1, 1))
par(mai = c(.8, .8, .2, .2))
plot(varImp_0,
decreasing = TRUE,
main = "Variable importance in complex model",
ylab = "variable importance"
)
# summarize the model details - not printed
# summary(smile__tree_model_0$finalModel)
# visualize the tree using the rattle package
fancyRpartPlot(smile__tree_model_0$finalModel)
# density plot of accuracy measurements, check with the resample data
trellis.par.set(caretTheme())
densityplot(smile__tree_model_0, pch = "|")
smile__tree_model_0$resample
## Accuracy Kappa Resample
## 1 0.5151515 0.03649635 Fold02
## 2 0.6363636 0.27737226 Fold01
## 3 0.5757576 0.15693431 Fold03
## 4 0.5757576 0.15073529 Fold06
## 5 0.5882353 0.17647059 Fold05
## 6 0.5000000 0.00000000 Fold04
## 7 0.5454545 0.10163339 Fold07
## 8 0.5882353 0.17647059 Fold10
## 9 0.3750000 -0.25000000 Fold09
## 10 0.4705882 -0.05882353 Fold08
# model 0: complete model with centering and scaling - outcome does not change
set.seed(1973)
smile__tree_model_0.0.1 <- train(smile_type ~ .,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10),
preProcess = c("center", "scale")
)
smile__tree_model_0.0.1$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5370544 0.07672893 0.07505544 0.1510397
## 2 0.05252525 0.5369652 0.07274720 0.07414327 0.1465495
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
varImp_0.0.1 <- varImp(smile__tree_model_0.0.1)
varImp_0.0.1
## rpart variable importance
##
## only 20 most important variables shown (out of 32)
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 65.43
## AU25_r_mean 63.30
## apex_mean 59.99
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## amplitude_mean 43.03
## lip_mean 43.03
## AU09_r_mean 40.62
## stage_mean 24.72
## AU14_r_mean 21.58
## AU20_r_mean 15.81
## AU15_r_mean 0.00
## age 0.00
## onset_mean 0.00
## AU23_r_mean 0.00
## AU12_r_mean 0.00
## pose_Ry_mean 0.00
## `gendermale ` 0.00
## AU04_r_mean 0.00
par(mfrow = c(1, 1))
par(mai = c(.8, .8, .2, .2))
plot(varImp_0,
decreasing = TRUE,
main = "Variable importance in complex model",
ylab = "variable importance"
)
# summary(smile__tree_model_0$finalModel)
# visualize the tree using the rattle package
fancyRpartPlot(smile__tree_model_0.0.1$finalModel)
# predict based on the test set and the model, store the model
set.seed(1973)
smile__tree_model_0_pred <- predict(smile__tree_model_0, tst_smile)
# summary of the prediction
summary(smile__tree_model_0_pred)
## spontaneous deliberate
## 79 63
# create a confusion matrix to evaluate the model
smile__tree_model_0_confM <- confusionMatrix(
smile__tree_model_0_pred,
tst_smile$smile_type
)
# print the confusion matrix
smile__tree_model_0_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 42 37
## deliberate 28 35
##
## Accuracy : 0.5423
## 95% CI : (0.4567, 0.6261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2251
##
## Kappa : 0.086
##
## Mcnemar's Test P-Value : 0.3211
##
## Sensitivity : 0.6000
## Specificity : 0.4861
## Pos Pred Value : 0.5316
## Neg Pred Value : 0.5556
## Prevalence : 0.4930
## Detection Rate : 0.2958
## Detection Prevalence : 0.5563
## Balanced Accuracy : 0.5431
##
## 'Positive' Class : spontaneous
##
# same way working for predicting boys and girls
set.seed(1973)
smile__tree_model_0.1_pred <- predict(smile__tree_model_0, tst_smile_boys)
summary(smile__tree_model_0.1_pred)
## spontaneous deliberate
## 35 42
smile__tree_model_0.1_confM <- confusionMatrix(
smile__tree_model_0.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_0.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 18 17
## deliberate 19 23
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.0616
##
## Mcnemar's Test P-Value : 0.8676
##
## Sensitivity : 0.4865
## Specificity : 0.5750
## Pos Pred Value : 0.5143
## Neg Pred Value : 0.5476
## Prevalence : 0.4805
## Detection Rate : 0.2338
## Detection Prevalence : 0.4545
## Balanced Accuracy : 0.5307
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_0.2_pred <- predict(smile__tree_model_0, tst_smile_girls)
summary(smile__tree_model_0.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_0.2_confM <- confusionMatrix(
smile__tree_model_0.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_0.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 9 12
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.06332
##
## Sensitivity : 0.7273
## Specificity : 0.3750
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.5511
##
## 'Positive' Class : spontaneous
##
# model 1 onset-apex-offset
set.seed(1973)
smile__tree_model_1 <- train(smile_type ~ onset_mean + offset_mean + apex_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_1$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.05757576 0.5402184 0.08170716 0.07437569 0.1470887
## 2 0.06262626 0.5373663 0.07391236 0.07093887 0.1393274
## 3 0.06666667 0.5165107 0.03871962 0.06565946 0.1244773
varImp(smile__tree_model_1)
## rpart variable importance
##
## Overall
## apex_mean 100.00
## offset_mean 51.79
## onset_mean 0.00
# summary(smile__tree_model_1$finalModel)
fancyRpartPlot(smile__tree_model_1$finalModel)
smile__tree_model_1_pred <- predict(smile__tree_model_1, tst_smile)
summary(smile__tree_model_1_pred)
## spontaneous deliberate
## 28 114
smile__tree_model_1_confM <- confusionMatrix(
smile__tree_model_1_pred,
tst_smile$smile_type
)
smile__tree_model_1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 12
## deliberate 54 60
##
## Accuracy : 0.5352
## 95% CI : (0.4497, 0.6193)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2786
##
## Kappa : 0.0624
##
## Mcnemar's Test P-Value : 4.494e-07
##
## Sensitivity : 0.2286
## Specificity : 0.8333
## Pos Pred Value : 0.5714
## Neg Pred Value : 0.5263
## Prevalence : 0.4930
## Detection Rate : 0.1127
## Detection Prevalence : 0.1972
## Balanced Accuracy : 0.5310
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_1.1_pred <- predict(smile__tree_model_1, tst_smile_boys)
summary(smile__tree_model_1.1_pred)
## spontaneous deliberate
## 17 60
smile__tree_model_1.1_confM <- confusionMatrix(
smile__tree_model_1.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_1.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 10 7
## deliberate 27 33
##
## Accuracy : 0.5584
## 95% CI : (0.4407, 0.6716)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.28475
##
## Kappa : 0.0972
##
## Mcnemar's Test P-Value : 0.00112
##
## Sensitivity : 0.2703
## Specificity : 0.8250
## Pos Pred Value : 0.5882
## Neg Pred Value : 0.5500
## Prevalence : 0.4805
## Detection Rate : 0.1299
## Detection Prevalence : 0.2208
## Balanced Accuracy : 0.5476
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_1.2_pred <- predict(smile__tree_model_1, tst_smile_girls)
summary(smile__tree_model_1.2_pred)
## spontaneous deliberate
## 11 54
smile__tree_model_1.2_confM <- confusionMatrix(
smile__tree_model_1.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_1.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 6 5
## deliberate 27 27
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495482
##
## Kappa : 0.0253
##
## Mcnemar's Test P-Value : 0.0002054
##
## Sensitivity : 0.18182
## Specificity : 0.84375
## Pos Pred Value : 0.54545
## Neg Pred Value : 0.50000
## Prevalence : 0.50769
## Detection Rate : 0.09231
## Detection Prevalence : 0.16923
## Balanced Accuracy : 0.51278
##
## 'Positive' Class : spontaneous
##
# model 1A onset
set.seed(1973)
smile__tree_model_1A <- train(smile_type ~ onset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_1A$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01515152 0.4710060 -0.055638375 0.08058915 0.15999168
## 2 0.02424242 0.4741254 -0.051455995 0.05907009 0.11360546
## 3 0.04242424 0.4986631 -0.005634066 0.04641479 0.08338809
# summary(smile__tree_model_1A$finalModel)
smile__tree_model_1A_pred <- predict(smile__tree_model_1A, tst_smile)
summary(smile__tree_model_1A_pred)
## spontaneous deliberate
## 0 142
smile__tree_model_1A_confM <- confusionMatrix(
smile__tree_model_1A_pred,
tst_smile$smile_type
)
smile__tree_model_1A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 70 72
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.000
## Specificity : 1.000
## Pos Pred Value : NaN
## Neg Pred Value : 0.507
## Prevalence : 0.493
## Detection Rate : 0.000
## Detection Prevalence : 0.000
## Balanced Accuracy : 0.500
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_1A.1_pred <- predict(smile__tree_model_1A, tst_smile_boys)
summary(smile__tree_model_1A.1_pred)
## spontaneous deliberate
## 0 77
smile__tree_model_1A.1_confM <- confusionMatrix(
smile__tree_model_1A.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_1A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 37 40
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 3.252e-09
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.5195
## Prevalence : 0.4805
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_1A.2_pred <- predict(smile__tree_model_1A, tst_smile_girls)
summary(smile__tree_model_1A.2_pred)
## spontaneous deliberate
## 0 65
smile__tree_model_1A.2_confM <- confusionMatrix(
smile__tree_model_1A.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_1A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 33 32
##
## Accuracy : 0.4923
## 95% CI : (0.366, 0.6193)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.6452
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 2.54e-08
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.4923
## Prevalence : 0.5077
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
# model 1B apex
set.seed(1973)
smile__tree_model_1B <- train(smile_type ~ apex_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_1B$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01818182 0.4931540 -0.01455893 0.06234358 0.1266862
## 2 0.02121212 0.4928866 -0.01707760 0.06605422 0.1333608
## 3 0.11515152 0.4866310 -0.03605245 0.03749912 0.0690707
# summary(smile__tree_model_1B$finalModel)
smile__tree_model_1B_pred <- predict(smile__tree_model_1B, tst_smile)
summary(smile__tree_model_1B_pred)
## spontaneous deliberate
## 23 119
smile__tree_model_1B_confM <- confusionMatrix(
smile__tree_model_1B_pred,
tst_smile$smile_type
)
smile__tree_model_1B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 10
## deliberate 57 62
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.0473
##
## Mcnemar's Test P-Value : 1.912e-08
##
## Sensitivity : 0.18571
## Specificity : 0.86111
## Pos Pred Value : 0.56522
## Neg Pred Value : 0.52101
## Prevalence : 0.49296
## Detection Rate : 0.09155
## Detection Prevalence : 0.16197
## Balanced Accuracy : 0.52341
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_1B.1_pred <- predict(smile__tree_model_1B, tst_smile_boys)
summary(smile__tree_model_1B.1_pred)
## spontaneous deliberate
## 14 63
smile__tree_model_1B.1_confM <- confusionMatrix(
smile__tree_model_1B.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_1B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 9 5
## deliberate 28 35
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.2125930
##
## Kappa : 0.1211
##
## Mcnemar's Test P-Value : 0.0001283
##
## Sensitivity : 0.2432
## Specificity : 0.8750
## Pos Pred Value : 0.6429
## Neg Pred Value : 0.5556
## Prevalence : 0.4805
## Detection Rate : 0.1169
## Detection Prevalence : 0.1818
## Balanced Accuracy : 0.5591
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_1B.2_pred <- predict(smile__tree_model_1B, tst_smile_girls)
summary(smile__tree_model_1B.2_pred)
## spontaneous deliberate
## 9 56
smile__tree_model_1B.2_confM <- confusionMatrix(
smile__tree_model_1B.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_1B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 4 5
## deliberate 29 27
##
## Accuracy : 0.4769
## 95% CI : (0.3515, 0.6046)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.7324
##
## Kappa : -0.0346
##
## Mcnemar's Test P-Value : 7.998e-05
##
## Sensitivity : 0.12121
## Specificity : 0.84375
## Pos Pred Value : 0.44444
## Neg Pred Value : 0.48214
## Prevalence : 0.50769
## Detection Rate : 0.06154
## Detection Prevalence : 0.13846
## Balanced Accuracy : 0.48248
##
## 'Positive' Class : spontaneous
##
# model 1C offset
set.seed(1973)
smile__tree_model_1C <- train(smile_type ~ offset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_1C$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01515152 0.4804924 -0.035544289 0.08003429 0.15971621
## 2 0.02424242 0.4951036 -0.005294115 0.07130717 0.14356728
## 3 0.06666667 0.5014316 0.016202110 0.03203681 0.05735123
# summary(smile__tree_model_1C$finalModel)
smile__tree_model_1C_pred <- predict(smile__tree_model_1C, tst_smile)
summary(smile__tree_model_1C_pred)
## spontaneous deliberate
## 0 142
smile__tree_model_1C_confM <- confusionMatrix(
smile__tree_model_1C_pred,
tst_smile$smile_type
)
smile__tree_model_1C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 70 72
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.000
## Specificity : 1.000
## Pos Pred Value : NaN
## Neg Pred Value : 0.507
## Prevalence : 0.493
## Detection Rate : 0.000
## Detection Prevalence : 0.000
## Balanced Accuracy : 0.500
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_1C.1_pred <- predict(smile__tree_model_1C, tst_smile_boys)
summary(smile__tree_model_1C.1_pred)
## spontaneous deliberate
## 0 77
smile__tree_model_1C.1_confM <- confusionMatrix(
smile__tree_model_1C.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_1C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 37 40
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 3.252e-09
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.5195
## Prevalence : 0.4805
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_1C.2_pred <- predict(smile__tree_model_1C, tst_smile_girls)
summary(smile__tree_model_1C.2_pred)
## spontaneous deliberate
## 0 65
smile__tree_model_1C.2_confM <- confusionMatrix(
smile__tree_model_1C.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_1C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 33 32
##
## Accuracy : 0.4923
## 95% CI : (0.366, 0.6193)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.6452
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 2.54e-08
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.4923
## Prevalence : 0.5077
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
# model 2 complete excluding subject and age info
set.seed(1973)
smile__tree_model_2 <- train(smile_type ~ . - subject - age,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_2$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5458779 0.09437598 0.07540099 0.1513876
## 2 0.05252525 0.5369652 0.07274720 0.07414327 0.1465495
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
varImp(smile__tree_model_2)
## rpart variable importance
##
## only 20 most important variables shown (out of 30)
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 65.43
## AU25_r_mean 63.30
## apex_mean 59.99
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## amplitude_mean 43.03
## lip_mean 43.03
## AU09_r_mean 40.62
## stage_mean 24.72
## AU14_r_mean 21.58
## AU20_r_mean 15.81
## AU23_r_mean 0.00
## gaze_angle_y_mean 0.00
## eye_mean 0.00
## AU17_r_mean 0.00
## AU02_r_mean 0.00
## gaze_angle_x_mean 0.00
## AU07_r_mean 0.00
## `gendermale ` 0.00
# summary(smile__tree_model_2$finalModel)
fancyRpartPlot(smile__tree_model_2$finalModel)
smile__tree_model_2_pred <- predict(smile__tree_model_2, tst_smile)
summary(smile__tree_model_2_pred)
## spontaneous deliberate
## 79 63
smile__tree_model_2_confM <- confusionMatrix(
smile__tree_model_2_pred,
tst_smile$smile_type
)
smile__tree_model_2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 42 37
## deliberate 28 35
##
## Accuracy : 0.5423
## 95% CI : (0.4567, 0.6261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2251
##
## Kappa : 0.086
##
## Mcnemar's Test P-Value : 0.3211
##
## Sensitivity : 0.6000
## Specificity : 0.4861
## Pos Pred Value : 0.5316
## Neg Pred Value : 0.5556
## Prevalence : 0.4930
## Detection Rate : 0.2958
## Detection Prevalence : 0.5563
## Balanced Accuracy : 0.5431
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_2.1_pred <- predict(smile__tree_model_2, tst_smile_boys)
summary(smile__tree_model_2.1_pred)
## spontaneous deliberate
## 35 42
smile__tree_model_2.1_confM <- confusionMatrix(
smile__tree_model_2.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_2.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 18 17
## deliberate 19 23
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.0616
##
## Mcnemar's Test P-Value : 0.8676
##
## Sensitivity : 0.4865
## Specificity : 0.5750
## Pos Pred Value : 0.5143
## Neg Pred Value : 0.5476
## Prevalence : 0.4805
## Detection Rate : 0.2338
## Detection Prevalence : 0.4545
## Balanced Accuracy : 0.5307
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_2.2_pred <- predict(smile__tree_model_2, tst_smile_girls)
summary(smile__tree_model_2.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_2.2_confM <- confusionMatrix(
smile__tree_model_2.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_2.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 9 12
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.06332
##
## Sensitivity : 0.7273
## Specificity : 0.3750
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.5511
##
## 'Positive' Class : spontaneous
##
# model 3 complete lip and eye features
set.seed(1973)
smile__tree_model_3 <- train(smile_type ~ lip_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_3$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01515152 0.5526738 0.10440266 0.09180319 0.18217867
## 2 0.06666667 0.5892435 0.17469528 0.10479850 0.21101409
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
varImp(smile__tree_model_3)
## rpart variable importance
##
## Overall
## eye_mean 100
## lip_mean 0
# summary(smile__tree_model_3$finalModel)
fancyRpartPlot(smile__tree_model_3$finalModel)
smile__tree_model_3_pred <- predict(smile__tree_model_3, tst_smile)
summary(smile__tree_model_3_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_3_confM <- confusionMatrix(
smile__tree_model_3_pred,
tst_smile$smile_type
)
smile__tree_model_3_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 45 56
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.076642
##
## Kappa : 0.1357
##
## Mcnemar's Test P-Value : 0.000337
##
## Sensitivity : 0.3571
## Specificity : 0.7778
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.5545
## Prevalence : 0.4930
## Detection Rate : 0.1761
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5675
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_3.1_pred <- predict(smile__tree_model_3, tst_smile_boys)
summary(smile__tree_model_3.1_pred)
## spontaneous deliberate
## 22 55
smile__tree_model_3.1_confM <- confusionMatrix(
smile__tree_model_3.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_3.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 9
## deliberate 24 31
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.21259
##
## Kappa : 0.1283
##
## Mcnemar's Test P-Value : 0.01481
##
## Sensitivity : 0.3514
## Specificity : 0.7750
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.5636
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2857
## Balanced Accuracy : 0.5632
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_3.2_pred <- predict(smile__tree_model_3, tst_smile_girls)
summary(smile__tree_model_3.2_pred)
## spontaneous deliberate
## 19 46
smile__tree_model_3.2_confM <- confusionMatrix(
smile__tree_model_3.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_3.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 7
## deliberate 21 25
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.19273
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.01402
##
## Sensitivity : 0.3636
## Specificity : 0.7812
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.5435
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2923
## Balanced Accuracy : 0.5724
##
## 'Positive' Class : spontaneous
##
# model 3A lip
set.seed(1973)
smile__tree_model_3A <- train(smile_type ~ lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_3A$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02121212 0.4811943 -0.03738362 0.09195689 0.18310126
## 2 0.02424242 0.4667725 -0.06529158 0.09635085 0.19215722
## 3 0.09090909 0.4836898 -0.03924734 0.04451978 0.08362762
# summary(smile__tree_model_3A$finalModel)
smile__tree_model_3A_pred <- predict(smile__tree_model_3A, tst_smile)
summary(smile__tree_model_3A_pred)
## spontaneous deliberate
## 0 142
smile__tree_model_3A_confM <- confusionMatrix(
smile__tree_model_3A_pred,
tst_smile$smile_type
)
smile__tree_model_3A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 70 72
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.000
## Specificity : 1.000
## Pos Pred Value : NaN
## Neg Pred Value : 0.507
## Prevalence : 0.493
## Detection Rate : 0.000
## Detection Prevalence : 0.000
## Balanced Accuracy : 0.500
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_3A.1_pred <- predict(smile__tree_model_3A, tst_smile_boys)
summary(smile__tree_model_3A.1_pred)
## spontaneous deliberate
## 0 77
smile__tree_model_3A.1_confM <- confusionMatrix(
smile__tree_model_3A.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_3A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 37 40
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 3.252e-09
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.5195
## Prevalence : 0.4805
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_3A.2_pred <- predict(smile__tree_model_3A, tst_smile_girls)
summary(smile__tree_model_3A.2_pred)
## spontaneous deliberate
## 0 65
smile__tree_model_3A.2_confM <- confusionMatrix(
smile__tree_model_3A.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_3A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 33 32
##
## Accuracy : 0.4923
## 95% CI : (0.366, 0.6193)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.6452
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 2.54e-08
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.4923
## Prevalence : 0.5077
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
# model 3B eye
set.seed(1973)
smile__tree_model_3B <- train(smile_type ~ eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_3B$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5889706 0.17587312 0.07422087 0.14742791
## 2 0.06666667 0.5892435 0.17469528 0.10479850 0.21101409
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
# summary(smile__tree_model_3B$finalModel)
smile__tree_model_3B_pred <- predict(smile__tree_model_3B, tst_smile)
summary(smile__tree_model_3B_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_3B_confM <- confusionMatrix(
smile__tree_model_3B_pred,
tst_smile$smile_type
)
smile__tree_model_3B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 45 56
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.076642
##
## Kappa : 0.1357
##
## Mcnemar's Test P-Value : 0.000337
##
## Sensitivity : 0.3571
## Specificity : 0.7778
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.5545
## Prevalence : 0.4930
## Detection Rate : 0.1761
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5675
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_3B.1_pred <- predict(smile__tree_model_3B, tst_smile_boys)
summary(smile__tree_model_3B.1_pred)
## spontaneous deliberate
## 22 55
smile__tree_model_3B.1_confM <- confusionMatrix(
smile__tree_model_3B.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_3B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 9
## deliberate 24 31
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.21259
##
## Kappa : 0.1283
##
## Mcnemar's Test P-Value : 0.01481
##
## Sensitivity : 0.3514
## Specificity : 0.7750
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.5636
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2857
## Balanced Accuracy : 0.5632
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_3B.2_pred <- predict(smile__tree_model_3B, tst_smile_girls)
summary(smile__tree_model_3B.2_pred)
## spontaneous deliberate
## 19 46
smile__tree_model_3B.2_confM <- confusionMatrix(
smile__tree_model_3B.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_3B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 7
## deliberate 21 25
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.19273
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.01402
##
## Sensitivity : 0.3636
## Specificity : 0.7812
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.5435
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2923
## Balanced Accuracy : 0.5724
##
## 'Positive' Class : spontaneous
##
# model 3C complete lip, amplitude, and eye features
set.seed(1973)
smile__tree_model_3C <- train(smile_type ~ lip_mean + eye_mean + amplitude_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_3C$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01515152 0.5526738 0.10440266 0.09180319 0.18217867
## 2 0.06666667 0.5892435 0.17469528 0.10479850 0.21101409
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
varImp(smile__tree_model_3C)
## rpart variable importance
##
## Overall
## eye_mean 100
## amplitude_mean 0
## lip_mean 0
# summary(smile__tree_model_3C$finalModel)
fancyRpartPlot(smile__tree_model_3C$finalModel)
smile__tree_model_3C_pred <- predict(smile__tree_model_3C, tst_smile)
summary(smile__tree_model_3C_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_3C_confM <- confusionMatrix(
smile__tree_model_3C_pred,
tst_smile$smile_type
)
smile__tree_model_3C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 45 56
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.076642
##
## Kappa : 0.1357
##
## Mcnemar's Test P-Value : 0.000337
##
## Sensitivity : 0.3571
## Specificity : 0.7778
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.5545
## Prevalence : 0.4930
## Detection Rate : 0.1761
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5675
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_3C.1_pred <- predict(smile__tree_model_3C, tst_smile_boys)
summary(smile__tree_model_3C.1_pred)
## spontaneous deliberate
## 22 55
smile__tree_model_3C.1_confM <- confusionMatrix(
smile__tree_model_3C.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_3C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 9
## deliberate 24 31
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.21259
##
## Kappa : 0.1283
##
## Mcnemar's Test P-Value : 0.01481
##
## Sensitivity : 0.3514
## Specificity : 0.7750
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.5636
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2857
## Balanced Accuracy : 0.5632
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_3C.2_pred <- predict(smile__tree_model_3C, tst_smile_girls)
summary(smile__tree_model_3C.2_pred)
## spontaneous deliberate
## 19 46
smile__tree_model_3C.2_confM <- confusionMatrix(
smile__tree_model_3C.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_3C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 7
## deliberate 21 25
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.19273
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.01402
##
## Sensitivity : 0.3636
## Specificity : 0.7812
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.5435
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2923
## Balanced Accuracy : 0.5724
##
## 'Positive' Class : spontaneous
##
# model 4 AU features complete model
set.seed(1973)
smile__tree_model_4 <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5636141 0.12907456 0.08437636 0.1685849
## 2 0.05252525 0.5308155 0.05932530 0.07759515 0.1551832
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_4$coefnames
## [1] "AU01_r_mean" "AU02_r_mean" "AU04_r_mean" "AU05_r_mean" "AU06_r_mean"
## [6] "AU07_r_mean" "AU09_r_mean" "AU10_r_mean" "AU12_r_mean" "AU14_r_mean"
## [11] "AU15_r_mean" "AU17_r_mean" "AU20_r_mean" "AU23_r_mean" "AU25_r_mean"
## [16] "AU26_r_mean" "AU45_r_mean"
varImp(smile__tree_model_4)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 81.66
## AU25_r_mean 77.78
## AU09_r_mean 73.78
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## AU14_r_mean 21.58
## AU12_r_mean 0.00
## AU23_r_mean 0.00
## AU07_r_mean 0.00
## AU02_r_mean 0.00
## AU04_r_mean 0.00
## AU06_r_mean 0.00
## AU17_r_mean 0.00
## AU20_r_mean 0.00
## AU26_r_mean 0.00
## AU15_r_mean 0.00
# summary(smile__tree_model_4$finalModel)
fancyRpartPlot(smile__tree_model_4$finalModel)
smile__tree_model_4_pred <- predict(smile__tree_model_4, tst_smile)
summary(smile__tree_model_4_pred)
## spontaneous deliberate
## 84 58
smile__tree_model_4_confM <- confusionMatrix(
smile__tree_model_4_pred,
tst_smile$smile_type
)
smile__tree_model_4_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 40
## deliberate 26 32
##
## Accuracy : 0.5352
## 95% CI : (0.4497, 0.6193)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2786
##
## Kappa : 0.0728
##
## Mcnemar's Test P-Value : 0.1096
##
## Sensitivity : 0.6286
## Specificity : 0.4444
## Pos Pred Value : 0.5238
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5915
## Balanced Accuracy : 0.5365
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4.1_pred <- predict(smile__tree_model_4, tst_smile_boys)
summary(smile__tree_model_4.1_pred)
## spontaneous deliberate
## 40 37
smile__tree_model_4.1_confM <- confusionMatrix(
smile__tree_model_4.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 20
## deliberate 17 20
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0.0404
##
## Mcnemar's Test P-Value : 0.7423
##
## Sensitivity : 0.5405
## Specificity : 0.5000
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5405
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.5195
## Balanced Accuracy : 0.5203
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4.2_pred <- predict(smile__tree_model_4, tst_smile_girls)
summary(smile__tree_model_4.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_4.2_confM <- confusionMatrix(
smile__tree_model_4.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 9 12
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.06332
##
## Sensitivity : 0.7273
## Specificity : 0.3750
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.5511
##
## 'Positive' Class : spontaneous
##
# model 4A AU features happiness model
set.seed(1973)
smile__tree_model_4A <- train(smile_type ~ AU06_r_mean + AU12_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(
method = "cv",
number = 10
)
)
# density plot of accuracy measurements
trellis.par.set(caretTheme())
densityplot(smile__tree_model_4A, pch = "|")
smile__tree_model_4A$resample
## Accuracy Kappa Resample
## 1 0.4545455 -0.09191176 Fold02
## 2 0.5454545 0.09506399 Fold01
## 3 0.6060606 0.22702703 Fold03
## 4 0.4848485 -0.02935780 Fold06
## 5 0.5000000 0.00000000 Fold05
## 6 0.4117647 -0.17647059 Fold04
## 7 0.5757576 0.14126394 Fold07
## 8 0.5294118 0.05882353 Fold10
## 9 0.4375000 -0.12500000 Fold09
## 10 0.4705882 -0.05882353 Fold08
smile__tree_model_4A$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01969697 0.5015931 0.00406148 0.06213453 0.12621231
## 2 0.02828283 0.4866199 -0.02825439 0.04520997 0.08863008
## 3 0.10909091 0.4985684 -0.01151477 0.02486434 0.04647333
smile__tree_model_4A$coefnames
## [1] "AU06_r_mean" "AU12_r_mean"
varImp(smile__tree_model_4A)
## rpart variable importance
##
## Overall
## AU06_r_mean 100
## AU12_r_mean 0
# summary(smile__tree_model_4A$finalModel)
fancyRpartPlot(smile__tree_model_4A$finalModel,
caption = "Model 4A: AU06 + AU12"
)
smile__tree_model_4A_pred <- predict(smile__tree_model_4A, tst_smile)
summary(smile__tree_model_4A_pred)
## spontaneous deliberate
## 38 104
smile__tree_model_4A_confM <- confusionMatrix(
smile__tree_model_4A_pred,
tst_smile$smile_type
)
smile__tree_model_4A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 18
## deliberate 50 54
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.4008030
##
## Kappa : 0.0359
##
## Mcnemar's Test P-Value : 0.0001704
##
## Sensitivity : 0.2857
## Specificity : 0.7500
## Pos Pred Value : 0.5263
## Neg Pred Value : 0.5192
## Prevalence : 0.4930
## Detection Rate : 0.1408
## Detection Prevalence : 0.2676
## Balanced Accuracy : 0.5179
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4A.1_pred <- predict(smile__tree_model_4A, tst_smile_boys)
summary(smile__tree_model_4A.1_pred)
## spontaneous deliberate
## 23 54
smile__tree_model_4A.1_confM <- confusionMatrix(
smile__tree_model_4A.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 10
## deliberate 24 30
##
## Accuracy : 0.5584
## 95% CI : (0.4407, 0.6716)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.28475
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.02578
##
## Sensitivity : 0.3514
## Specificity : 0.7500
## Pos Pred Value : 0.5652
## Neg Pred Value : 0.5556
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2987
## Balanced Accuracy : 0.5507
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4A.2_pred <- predict(smile__tree_model_4A, tst_smile_girls)
summary(smile__tree_model_4A.2_pred)
## spontaneous deliberate
## 15 50
smile__tree_model_4A.2_confM <- confusionMatrix(
smile__tree_model_4A.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 7 8
## deliberate 26 24
##
## Accuracy : 0.4769
## 95% CI : (0.3515, 0.6046)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.732416
##
## Kappa : -0.0376
##
## Mcnemar's Test P-Value : 0.003551
##
## Sensitivity : 0.2121
## Specificity : 0.7500
## Pos Pred Value : 0.4667
## Neg Pred Value : 0.4800
## Prevalence : 0.5077
## Detection Rate : 0.1077
## Detection Prevalence : 0.2308
## Balanced Accuracy : 0.4811
##
## 'Positive' Class : spontaneous
##
# model 4B AU best model
set.seed(1973)
smile__tree_model_4B <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(
method = "cv",
number = 10
)
)
smile__tree_model_4B$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5547126 0.11324964 0.07576006 0.1500753
## 2 0.04545455 0.5364305 0.07335263 0.07508962 0.1507522
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_4B$coefnames
## [1] "AU01_r_mean" "AU09_r_mean" "AU10_r_mean" "AU25_r_mean" "AU45_r_mean"
varImp(smile__tree_model_4B)
## rpart variable importance
##
## Overall
## AU25_r_mean 100.00
## AU45_r_mean 65.35
## AU01_r_mean 47.05
## AU10_r_mean 15.19
## AU09_r_mean 0.00
# summary(smile__tree_model_4$finalModel)
fancyRpartPlot(smile__tree_model_4B$finalModel)
smile__tree_model_4B_pred <- predict(smile__tree_model_4B, tst_smile)
summary(smile__tree_model_4B_pred)
## spontaneous deliberate
## 89 53
smile__tree_model_4B_confM <- confusionMatrix(
smile__tree_model_4B_pred,
tst_smile$smile_type
)
smile__tree_model_4B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 49 40
## deliberate 21 32
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.07664
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.02119
##
## Sensitivity : 0.7000
## Specificity : 0.4444
## Pos Pred Value : 0.5506
## Neg Pred Value : 0.6038
## Prevalence : 0.4930
## Detection Rate : 0.3451
## Detection Prevalence : 0.6268
## Balanced Accuracy : 0.5722
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4B.1_pred <- predict(smile__tree_model_4B, tst_smile_boys)
summary(smile__tree_model_4B.1_pred)
## spontaneous deliberate
## 41 36
smile__tree_model_4B.1_confM <- confusionMatrix(
smile__tree_model_4B.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 18
## deliberate 14 22
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1709
##
## Mcnemar's Test P-Value : 0.5959
##
## Sensitivity : 0.6216
## Specificity : 0.5500
## Pos Pred Value : 0.5610
## Neg Pred Value : 0.6111
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5325
## Balanced Accuracy : 0.5858
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4B.2_pred <- predict(smile__tree_model_4B, tst_smile_girls)
summary(smile__tree_model_4B.2_pred)
## spontaneous deliberate
## 48 17
smile__tree_model_4B.2_confM <- confusionMatrix(
smile__tree_model_4B.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 22
## deliberate 7 10
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1011
##
## Mcnemar's Test P-Value : 0.00933
##
## Sensitivity : 0.7879
## Specificity : 0.3125
## Pos Pred Value : 0.5417
## Neg Pred Value : 0.5882
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.7385
## Balanced Accuracy : 0.5502
##
## 'Positive' Class : spontaneous
##
# model 4C AU happiness + blink
set.seed(1973)
smile__tree_model_4C <- train(smile_type ~ AU45_r_mean + AU06_r_mean +
AU12_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(
method = "cv",
number = 10
)
)
smile__tree_model_4C$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02424242 0.5797850 0.16016594 0.05522519 0.11234354
## 2 0.06060606 0.5464516 0.09717643 0.04388348 0.08896946
## 3 0.18787879 0.5223708 0.04901466 0.03672728 0.07788091
smile__tree_model_4C$coefnames
## [1] "AU45_r_mean" "AU06_r_mean" "AU12_r_mean"
varImp(smile__tree_model_4C)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU06_r_mean 31.92
## AU12_r_mean 0.00
# summary(smile__tree_model_4C$finalModel)
fancyRpartPlot(smile__tree_model_4C$finalModel)
smile__tree_model_4C_pred <- predict(smile__tree_model_4C, tst_smile)
summary(smile__tree_model_4C_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_4C_confM <- confusionMatrix(
smile__tree_model_4C_pred,
tst_smile$smile_type
)
smile__tree_model_4C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4C.1_pred <- predict(smile__tree_model_4C, tst_smile_boys)
summary(smile__tree_model_4C.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_4C.1_confM <- confusionMatrix(
smile__tree_model_4C.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4C.2_pred <- predict(smile__tree_model_4C, tst_smile_girls)
summary(smile__tree_model_4C.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_4C.2_confM <- confusionMatrix(
smile__tree_model_4C.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# model 4D AU45
set.seed(1973)
smile__tree_model_4D <- train(smile_type ~ AU45_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4D$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02121212 0.5617814 0.12519241 0.03886699 0.08002424
## 2 0.06060606 0.5464516 0.09717643 0.04388348 0.08896946
## 3 0.18787879 0.5223708 0.04901466 0.03672728 0.07788091
smile__tree_model_4D$coefnames
## [1] "AU45_r_mean"
# summary(smile__tree_model_4D$finalModel)
smile__tree_model_4D_pred <- predict(smile__tree_model_4D, tst_smile)
summary(smile__tree_model_4D_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_4D_confM <- confusionMatrix(
smile__tree_model_4D_pred,
tst_smile$smile_type
)
smile__tree_model_4D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4D.1_pred <- predict(smile__tree_model_4D, tst_smile_boys)
summary(smile__tree_model_4D.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_4D.1_confM <- confusionMatrix(
smile__tree_model_4D.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4D.2_pred <- predict(smile__tree_model_4D, tst_smile_girls)
summary(smile__tree_model_4D.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_4D.2_confM <- confusionMatrix(
smile__tree_model_4D.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# model 4E AU12
set.seed(1973)
smile__tree_model_4E <- train(smile_type ~ AU12_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4E$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01818182 0.5102496 0.020310865 0.08191055 0.16558148
## 2 0.03636364 0.5100713 0.019546074 0.05974664 0.11914570
## 3 0.06666667 0.4951872 -0.004623925 0.05086760 0.09953796
smile__tree_model_4E$coefnames
## [1] "AU12_r_mean"
# summary(smile__tree_model_4E$finalModel)
smile__tree_model_4E_pred <- predict(smile__tree_model_4E, tst_smile)
summary(smile__tree_model_4E_pred)
## spontaneous deliberate
## 58 84
smile__tree_model_4E_confM <- confusionMatrix(
smile__tree_model_4E_pred,
tst_smile$smile_type
)
smile__tree_model_4E_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 29 29
## deliberate 41 43
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0.0115
##
## Mcnemar's Test P-Value : 0.1886
##
## Sensitivity : 0.4143
## Specificity : 0.5972
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5119
## Prevalence : 0.4930
## Detection Rate : 0.2042
## Detection Prevalence : 0.4085
## Balanced Accuracy : 0.5058
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4E.1_pred <- predict(smile__tree_model_4E, tst_smile_boys)
summary(smile__tree_model_4E.1_pred)
## spontaneous deliberate
## 21 56
smile__tree_model_4E.1_confM <- confusionMatrix(
smile__tree_model_4E.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 8
## deliberate 24 32
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.15232
##
## Kappa : 0.1538
##
## Mcnemar's Test P-Value : 0.00801
##
## Sensitivity : 0.3514
## Specificity : 0.8000
## Pos Pred Value : 0.6190
## Neg Pred Value : 0.5714
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2727
## Balanced Accuracy : 0.5757
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4E.2_pred <- predict(smile__tree_model_4E, tst_smile_girls)
summary(smile__tree_model_4E.2_pred)
## spontaneous deliberate
## 37 28
smile__tree_model_4E.2_confM <- confusionMatrix(
smile__tree_model_4E.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 21
## deliberate 17 11
##
## Accuracy : 0.4154
## 95% CI : (0.2944, 0.5444)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.9468
##
## Kappa : -0.1717
##
## Mcnemar's Test P-Value : 0.6265
##
## Sensitivity : 0.4848
## Specificity : 0.3438
## Pos Pred Value : 0.4324
## Neg Pred Value : 0.3929
## Prevalence : 0.5077
## Detection Rate : 0.2462
## Detection Prevalence : 0.5692
## Balanced Accuracy : 0.4143
##
## 'Positive' Class : spontaneous
##
# model 4F AU06
set.seed(1973)
smile__tree_model_4F <- train(smile_type ~ AU06_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4F$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02020202 0.5135361 0.02820820 0.06572235 0.13344946
## 2 0.02828283 0.4896502 -0.02109638 0.05230700 0.10518620
## 3 0.10909091 0.4985684 -0.01151477 0.02486434 0.04647333
smile__tree_model_4F$coefnames
## [1] "AU06_r_mean"
# summary(smile__tree_model_4F$finalModel)
smile__tree_model_4F_pred <- predict(smile__tree_model_4F, tst_smile)
summary(smile__tree_model_4F_pred)
## spontaneous deliberate
## 47 95
smile__tree_model_4F_confM <- confusionMatrix(
smile__tree_model_4F_pred,
tst_smile$smile_type
)
smile__tree_model_4F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 25
## deliberate 48 47
##
## Accuracy : 0.4859
## 95% CI : (0.4013, 0.5712)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.72157
##
## Kappa : -0.0331
##
## Mcnemar's Test P-Value : 0.01003
##
## Sensitivity : 0.3143
## Specificity : 0.6528
## Pos Pred Value : 0.4681
## Neg Pred Value : 0.4947
## Prevalence : 0.4930
## Detection Rate : 0.1549
## Detection Prevalence : 0.3310
## Balanced Accuracy : 0.4835
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4F.1_pred <- predict(smile__tree_model_4F, tst_smile_boys)
summary(smile__tree_model_4F.1_pred)
## spontaneous deliberate
## 26 51
smile__tree_model_4F.1_confM <- confusionMatrix(
smile__tree_model_4F.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 13
## deliberate 24 27
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0.0266
##
## Mcnemar's Test P-Value : 0.1002
##
## Sensitivity : 0.3514
## Specificity : 0.6750
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5294
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.3377
## Balanced Accuracy : 0.5132
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4F.2_pred <- predict(smile__tree_model_4F, tst_smile_girls)
summary(smile__tree_model_4F.2_pred)
## spontaneous deliberate
## 21 44
smile__tree_model_4F.2_confM <- confusionMatrix(
smile__tree_model_4F.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 9 12
## deliberate 24 20
##
## Accuracy : 0.4462
## 95% CI : (0.3227, 0.5747)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.86792
##
## Kappa : -0.1017
##
## Mcnemar's Test P-Value : 0.06675
##
## Sensitivity : 0.2727
## Specificity : 0.6250
## Pos Pred Value : 0.4286
## Neg Pred Value : 0.4545
## Prevalence : 0.5077
## Detection Rate : 0.1385
## Detection Prevalence : 0.3231
## Balanced Accuracy : 0.4489
##
## 'Positive' Class : spontaneous
##
# model 4G AU10
set.seed(1973)
smile__tree_model_4G <- train(smile_type ~ AU10_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4G$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01212121 0.5520611 0.10499803 0.06372484 0.12732122
## 2 0.01414141 0.5253175 0.05036962 0.03849244 0.07670847
## 3 0.22424242 0.5437667 0.08478626 0.04730443 0.09468424
smile__tree_model_4G$coefnames
## [1] "AU10_r_mean"
# summary(smile__tree_model_4G$finalModel)
smile__tree_model_4G_pred <- predict(smile__tree_model_4G, tst_smile)
summary(smile__tree_model_4G_pred)
## spontaneous deliberate
## 76 66
smile__tree_model_4G_confM <- confusionMatrix(
smile__tree_model_4G_pred,
tst_smile$smile_type
)
smile__tree_model_4G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 38 38
## deliberate 32 34
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0.0151
##
## Mcnemar's Test P-Value : 0.5501
##
## Sensitivity : 0.5429
## Specificity : 0.4722
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5152
## Prevalence : 0.4930
## Detection Rate : 0.2676
## Detection Prevalence : 0.5352
## Balanced Accuracy : 0.5075
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4G.1_pred <- predict(smile__tree_model_4G, tst_smile_boys)
summary(smile__tree_model_4G.1_pred)
## spontaneous deliberate
## 32 45
smile__tree_model_4G.1_confM <- confusionMatrix(
smile__tree_model_4G.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 16
## deliberate 21 24
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0.0326
##
## Mcnemar's Test P-Value : 0.5108
##
## Sensitivity : 0.4324
## Specificity : 0.6000
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5333
## Prevalence : 0.4805
## Detection Rate : 0.2078
## Detection Prevalence : 0.4156
## Balanced Accuracy : 0.5162
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4G.2_pred <- predict(smile__tree_model_4G, tst_smile_girls)
summary(smile__tree_model_4G.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_4G.2_confM <- confusionMatrix(
smile__tree_model_4G.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 22
## deliberate 11 10
##
## Accuracy : 0.4923
## 95% CI : (0.366, 0.6193)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.64516
##
## Kappa : -0.0209
##
## Mcnemar's Test P-Value : 0.08172
##
## Sensitivity : 0.6667
## Specificity : 0.3125
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.4762
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.4896
##
## 'Positive' Class : spontaneous
##
# model 4H AU01
set.seed(1973)
smile__tree_model_4H <- train(smile_type ~ AU01_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4H$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01818182 0.5226437 0.04911567 0.09722360 0.1916064
## 2 0.03030303 0.5528743 0.10890407 0.09706592 0.1908838
## 3 0.14545455 0.5290775 0.06682929 0.09259357 0.1810430
smile__tree_model_4H$coefnames
## [1] "AU01_r_mean"
# summary(smile__tree_model_4H$finalModel)
smile__tree_model_4H_pred <- predict(smile__tree_model_4H, tst_smile)
summary(smile__tree_model_4H_pred)
## spontaneous deliberate
## 107 35
smile__tree_model_4H_confM <- confusionMatrix(
smile__tree_model_4H_pred,
tst_smile$smile_type
)
smile__tree_model_4H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 56 51
## deliberate 14 21
##
## Accuracy : 0.5423
## 95% CI : (0.4567, 0.6261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2251
##
## Kappa : 0.091
##
## Mcnemar's Test P-Value : 7.998e-06
##
## Sensitivity : 0.8000
## Specificity : 0.2917
## Pos Pred Value : 0.5234
## Neg Pred Value : 0.6000
## Prevalence : 0.4930
## Detection Rate : 0.3944
## Detection Prevalence : 0.7535
## Balanced Accuracy : 0.5458
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4H.1_pred <- predict(smile__tree_model_4H, tst_smile_boys)
summary(smile__tree_model_4H.1_pred)
## spontaneous deliberate
## 61 16
smile__tree_model_4H.1_confM <- confusionMatrix(
smile__tree_model_4H.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 32 29
## deliberate 5 11
##
## Accuracy : 0.5584
## 95% CI : (0.4407, 0.6716)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.2847
##
## Kappa : 0.1365
##
## Mcnemar's Test P-Value : 7.998e-05
##
## Sensitivity : 0.8649
## Specificity : 0.2750
## Pos Pred Value : 0.5246
## Neg Pred Value : 0.6875
## Prevalence : 0.4805
## Detection Rate : 0.4156
## Detection Prevalence : 0.7922
## Balanced Accuracy : 0.5699
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4H.2_pred <- predict(smile__tree_model_4H, tst_smile_girls)
summary(smile__tree_model_4H.2_pred)
## spontaneous deliberate
## 46 19
smile__tree_model_4H.2_confM <- confusionMatrix(
smile__tree_model_4H.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 22
## deliberate 9 10
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.45095
##
## Kappa : 0.04
##
## Mcnemar's Test P-Value : 0.03114
##
## Sensitivity : 0.7273
## Specificity : 0.3125
## Pos Pred Value : 0.5217
## Neg Pred Value : 0.5263
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.7077
## Balanced Accuracy : 0.5199
##
## 'Positive' Class : spontaneous
##
# model 4I AU25
set.seed(1973)
smile__tree_model_4I <- train(smile_type ~ AU25_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4I$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03333333 0.5639149 0.13205234 0.06160107 0.12264971
## 2 0.06666667 0.5612411 0.12788597 0.03840889 0.06994947
## 3 0.14545455 0.5076705 0.02165391 0.01935801 0.04079529
smile__tree_model_4I$coefnames
## [1] "AU25_r_mean"
# summary(smile__tree_model_4I$finalModel)
smile__tree_model_4I_pred <- predict(smile__tree_model_4I, tst_smile)
summary(smile__tree_model_4I_pred)
## spontaneous deliberate
## 122 20
smile__tree_model_4I_confM <- confusionMatrix(
smile__tree_model_4I_pred,
tst_smile$smile_type
)
smile__tree_model_4I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 65 57
## deliberate 5 15
##
## Accuracy : 0.5634
## 95% CI : (0.4777, 0.6464)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.1039
##
## Kappa : 0.1355
##
## Mcnemar's Test P-Value : 9.356e-11
##
## Sensitivity : 0.9286
## Specificity : 0.2083
## Pos Pred Value : 0.5328
## Neg Pred Value : 0.7500
## Prevalence : 0.4930
## Detection Rate : 0.4577
## Detection Prevalence : 0.8592
## Balanced Accuracy : 0.5685
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4I.1_pred <- predict(smile__tree_model_4I, tst_smile_boys)
summary(smile__tree_model_4I.1_pred)
## spontaneous deliberate
## 68 9
smile__tree_model_4I.1_confM <- confusionMatrix(
smile__tree_model_4I.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 35 33
## deliberate 2 7
##
## Accuracy : 0.5455
## 95% CI : (0.4279, 0.6594)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.3667
##
## Kappa : 0.1173
##
## Mcnemar's Test P-Value : 3.959e-07
##
## Sensitivity : 0.9459
## Specificity : 0.1750
## Pos Pred Value : 0.5147
## Neg Pred Value : 0.7778
## Prevalence : 0.4805
## Detection Rate : 0.4545
## Detection Prevalence : 0.8831
## Balanced Accuracy : 0.5605
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4I.2_pred <- predict(smile__tree_model_4I, tst_smile_girls)
summary(smile__tree_model_4I.2_pred)
## spontaneous deliberate
## 54 11
smile__tree_model_4I.2_confM <- confusionMatrix(
smile__tree_model_4I.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 30 24
## deliberate 3 8
##
## Accuracy : 0.5846
## 95% CI : (0.4556, 0.7056)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.1320123
##
## Kappa : 0.1607
##
## Mcnemar's Test P-Value : 0.0001186
##
## Sensitivity : 0.9091
## Specificity : 0.2500
## Pos Pred Value : 0.5556
## Neg Pred Value : 0.7273
## Prevalence : 0.5077
## Detection Rate : 0.4615
## Detection Prevalence : 0.8308
## Balanced Accuracy : 0.5795
##
## 'Positive' Class : spontaneous
##
# model 4J AU09
set.seed(1973)
smile__tree_model_4J <- train(smile_type ~ AU09_r_mean,
method = "rpart",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_4J$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01818182 0.5195131 0.04226591 0.06740442 0.13251259
## 2 0.02222222 0.5195131 0.04226591 0.06740442 0.13251259
## 3 0.16363636 0.5073028 0.01055068 0.02369084 0.04615194
smile__tree_model_4J$coefnames
## [1] "AU09_r_mean"
# summary(smile__tree_model_4J$finalModel)
smile__tree_model_4J_pred <- predict(smile__tree_model_4J, tst_smile)
summary(smile__tree_model_4J_pred)
## spontaneous deliberate
## 109 33
smile__tree_model_4J_confM <- confusionMatrix(
smile__tree_model_4J_pred,
tst_smile$smile_type
)
smile__tree_model_4J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 54 55
## deliberate 16 17
##
## Accuracy : 0.5
## 95% CI : (0.415, 0.585)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5995
##
## Kappa : 0.0075
##
## Mcnemar's Test P-Value : 6.49e-06
##
## Sensitivity : 0.7714
## Specificity : 0.2361
## Pos Pred Value : 0.4954
## Neg Pred Value : 0.5152
## Prevalence : 0.4930
## Detection Rate : 0.3803
## Detection Prevalence : 0.7676
## Balanced Accuracy : 0.5038
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_4J.1_pred <- predict(smile__tree_model_4J, tst_smile_boys)
summary(smile__tree_model_4J.1_pred)
## spontaneous deliberate
## 57 20
smile__tree_model_4J.1_confM <- confusionMatrix(
smile__tree_model_4J.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_4J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 27 30
## deliberate 10 10
##
## Accuracy : 0.4805
## 95% CI : (0.3652, 0.5974)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.787723
##
## Kappa : -0.0199
##
## Mcnemar's Test P-Value : 0.002663
##
## Sensitivity : 0.7297
## Specificity : 0.2500
## Pos Pred Value : 0.4737
## Neg Pred Value : 0.5000
## Prevalence : 0.4805
## Detection Rate : 0.3506
## Detection Prevalence : 0.7403
## Balanced Accuracy : 0.4899
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_4J.2_pred <- predict(smile__tree_model_4J, tst_smile_girls)
summary(smile__tree_model_4J.2_pred)
## spontaneous deliberate
## 52 13
smile__tree_model_4J.2_confM <- confusionMatrix(
smile__tree_model_4J.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_4J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 27 25
## deliberate 6 7
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.450951
##
## Kappa : 0.0373
##
## Mcnemar's Test P-Value : 0.001225
##
## Sensitivity : 0.8182
## Specificity : 0.2188
## Pos Pred Value : 0.5192
## Neg Pred Value : 0.5385
## Prevalence : 0.5077
## Detection Rate : 0.4154
## Detection Prevalence : 0.8000
## Balanced Accuracy : 0.5185
##
## 'Positive' Class : spontaneous
##
# model 5 head pose features
set.seed(1973)
smile__tree_model_5 <-
train(smile_type ~ pose_Rx_mean + pose_Ry_mean + pose_Rz_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_5$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02575758 0.4929590 -0.018628892 0.06773807 0.13394796
## 2 0.02727273 0.5048964 0.005640949 0.07089130 0.14078111
## 3 0.10909091 0.5015096 -0.002110352 0.02310174 0.03990447
smile__tree_model_5$coefnames
## [1] "pose_Rx_mean" "pose_Ry_mean" "pose_Rz_mean"
varImp(smile__tree_model_5)
## rpart variable importance
##
## Overall
## pose_Rz_mean 100.00
## pose_Ry_mean 19.25
## pose_Rx_mean 0.00
# summary(smile__tree_model_5$finalModel)
smile__tree_model_5_pred <- predict(smile__tree_model_5, tst_smile)
summary(smile__tree_model_5_pred)
## spontaneous deliberate
## 42 100
smile__tree_model_5_confM <- confusionMatrix(
smile__tree_model_5_pred,
tst_smile$smile_type
)
smile__tree_model_5_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 22
## deliberate 50 50
##
## Accuracy : 0.493
## 95% CI : (0.4081, 0.5781)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.662667
##
## Kappa : -0.02
##
## Mcnemar's Test P-Value : 0.001463
##
## Sensitivity : 0.2857
## Specificity : 0.6944
## Pos Pred Value : 0.4762
## Neg Pred Value : 0.5000
## Prevalence : 0.4930
## Detection Rate : 0.1408
## Detection Prevalence : 0.2958
## Balanced Accuracy : 0.4901
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_5.1_pred <- predict(smile__tree_model_5, tst_smile_boys)
summary(smile__tree_model_5.1_pred)
## spontaneous deliberate
## 19 58
smile__tree_model_5.1_confM <- confusionMatrix(
smile__tree_model_5.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_5.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 8 11
## deliberate 29 29
##
## Accuracy : 0.4805
## 95% CI : (0.3652, 0.5974)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.78772
##
## Kappa : -0.0599
##
## Mcnemar's Test P-Value : 0.00719
##
## Sensitivity : 0.2162
## Specificity : 0.7250
## Pos Pred Value : 0.4211
## Neg Pred Value : 0.5000
## Prevalence : 0.4805
## Detection Rate : 0.1039
## Detection Prevalence : 0.2468
## Balanced Accuracy : 0.4706
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_5.2_pred <- predict(smile__tree_model_5, tst_smile_girls)
summary(smile__tree_model_5.2_pred)
## spontaneous deliberate
## 23 42
smile__tree_model_5.2_confM <- confusionMatrix(
smile__tree_model_5.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_5.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 11
## deliberate 21 21
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0198
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.3636
## Specificity : 0.6562
## Pos Pred Value : 0.5217
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.3538
## Balanced Accuracy : 0.5099
##
## 'Positive' Class : spontaneous
##
# model 5A gaze features
set.seed(1973)
smile__tree_model_5A <-
train(smile_type ~ gaze_angle_x_mean + gaze_angle_y_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_5A$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03939394 0.4440062 -0.10718904 0.08008631 0.1553984
## 2 0.04242424 0.4590686 -0.08001649 0.08664907 0.1694805
## 3 0.09090909 0.4810160 -0.04705882 0.05472730 0.1030112
smile__tree_model_5A$coefnames
## [1] "gaze_angle_x_mean" "gaze_angle_y_mean"
varImp(smile__tree_model_5A)
## rpart variable importance
##
## Overall
## gaze_angle_x_mean NaN
## gaze_angle_y_mean NaN
# summary(smile__tree_model_5A$finalModel)
smile__tree_model_5A_pred <- predict(smile__tree_model_5A, tst_smile)
summary(smile__tree_model_5A_pred)
## spontaneous deliberate
## 0 142
smile__tree_model_5A_confM <- confusionMatrix(
smile__tree_model_5A_pred,
tst_smile$smile_type
)
smile__tree_model_5A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 70 72
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.000
## Specificity : 1.000
## Pos Pred Value : NaN
## Neg Pred Value : 0.507
## Prevalence : 0.493
## Detection Rate : 0.000
## Detection Prevalence : 0.000
## Balanced Accuracy : 0.500
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_5A.1_pred <- predict(smile__tree_model_5A, tst_smile_boys)
summary(smile__tree_model_5A.1_pred)
## spontaneous deliberate
## 0 77
smile__tree_model_5A.1_confM <- confusionMatrix(
smile__tree_model_5A.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_5A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 37 40
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 3.252e-09
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.5195
## Prevalence : 0.4805
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_5A.2_pred <- predict(smile__tree_model_5A, tst_smile_girls)
summary(smile__tree_model_5A.2_pred)
## spontaneous deliberate
## 0 65
smile__tree_model_5A.2_confM <- confusionMatrix(
smile__tree_model_5A.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_5A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 33 32
##
## Accuracy : 0.4923
## 95% CI : (0.366, 0.6193)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.6452
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 2.54e-08
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.4923
## Prevalence : 0.5077
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
# model 5B headpose + gaze features
set.seed(1973)
smile__tree_model_5B <-
train(smile_type ~ pose_Rx_mean + pose_Ry_mean + pose_Rz_mean +
gaze_angle_x_mean + gaze_angle_y_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_5B$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.01969697 0.4623719 -0.073073994 0.07260347 0.14825324
## 2 0.02575758 0.4686999 -0.064517461 0.07874648 0.16003545
## 3 0.10909091 0.5015096 -0.002110352 0.02310174 0.03990447
smile__tree_model_5B$coefnames
## [1] "pose_Rx_mean" "pose_Ry_mean" "pose_Rz_mean"
## [4] "gaze_angle_x_mean" "gaze_angle_y_mean"
varImp(smile__tree_model_5B)
## rpart variable importance
##
## Overall
## gaze_angle_y_mean NaN
## gaze_angle_x_mean NaN
## pose_Rx_mean NaN
## pose_Ry_mean NaN
## pose_Rz_mean NaN
# summary(smile__tree_model_5B$finalModel)
smile__tree_model_5B_pred <- predict(smile__tree_model_5B, tst_smile)
summary(smile__tree_model_5B_pred)
## spontaneous deliberate
## 0 142
smile__tree_model_5B_confM <- confusionMatrix(
smile__tree_model_5B_pred,
tst_smile$smile_type
)
smile__tree_model_5B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 70 72
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.000
## Specificity : 1.000
## Pos Pred Value : NaN
## Neg Pred Value : 0.507
## Prevalence : 0.493
## Detection Rate : 0.000
## Detection Prevalence : 0.000
## Balanced Accuracy : 0.500
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_5B.1_pred <- predict(smile__tree_model_5B, tst_smile_boys)
summary(smile__tree_model_5B.1_pred)
## spontaneous deliberate
## 0 77
smile__tree_model_5B.1_confM <- confusionMatrix(
smile__tree_model_5B.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_5B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 37 40
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 3.252e-09
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.5195
## Prevalence : 0.4805
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_5B.2_pred <- predict(smile__tree_model_5B, tst_smile_girls)
summary(smile__tree_model_5B.2_pred)
## spontaneous deliberate
## 0 65
smile__tree_model_5B.2_confM <- confusionMatrix(
smile__tree_model_5B.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_5B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 33 32
##
## Accuracy : 0.4923
## 95% CI : (0.366, 0.6193)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.6452
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 2.54e-08
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.4923
## Prevalence : 0.5077
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
# model 6 dynamics and movement
set.seed(1973)
smile__tree_model_6 <-
train(smile_type ~ onset_mean + apex_mean + offset_mean + eye_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5492090 0.09642256 0.09355662 0.18550882
## 2 0.06666667 0.5314728 0.06480625 0.09078317 0.17739898
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6$coefnames
## [1] "onset_mean" "apex_mean" "offset_mean" "eye_mean" "lip_mean"
varImp(smile__tree_model_6)
## rpart variable importance
##
## Overall
## eye_mean 100.00
## offset_mean 77.33
## onset_mean 43.18
## apex_mean 19.36
## lip_mean 0.00
# summary(smile__tree_model_6$finalModel)
fancyRpartPlot(smile__tree_model_6$finalModel)
smile__tree_model_6_pred <- predict(smile__tree_model_6, tst_smile)
summary(smile__tree_model_6_pred)
## spontaneous deliberate
## 88 54
smile__tree_model_6_confM <- confusionMatrix(
smile__tree_model_6_pred,
tst_smile$smile_type
)
smile__tree_model_6_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 45 43
## deliberate 25 29
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.40080
##
## Kappa : 0.0455
##
## Mcnemar's Test P-Value : 0.03925
##
## Sensitivity : 0.6429
## Specificity : 0.4028
## Pos Pred Value : 0.5114
## Neg Pred Value : 0.5370
## Prevalence : 0.4930
## Detection Rate : 0.3169
## Detection Prevalence : 0.6197
## Balanced Accuracy : 0.5228
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6.1_pred <- predict(smile__tree_model_6, tst_smile_boys)
summary(smile__tree_model_6.1_pred)
## spontaneous deliberate
## 52 25
smile__tree_model_6.1_confM <- confusionMatrix(
smile__tree_model_6.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 26
## deliberate 11 14
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.54593
##
## Kappa : 0.0519
##
## Mcnemar's Test P-Value : 0.02136
##
## Sensitivity : 0.7027
## Specificity : 0.3500
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5600
## Prevalence : 0.4805
## Detection Rate : 0.3377
## Detection Prevalence : 0.6753
## Balanced Accuracy : 0.5264
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6.2_pred <- predict(smile__tree_model_6, tst_smile_girls)
summary(smile__tree_model_6.2_pred)
## spontaneous deliberate
## 36 29
smile__tree_model_6.2_confM <- confusionMatrix(
smile__tree_model_6.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 17
## deliberate 14 15
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0446
##
## Mcnemar's Test P-Value : 0.7194
##
## Sensitivity : 0.5758
## Specificity : 0.4688
## Pos Pred Value : 0.5278
## Neg Pred Value : 0.5172
## Prevalence : 0.5077
## Detection Rate : 0.2923
## Detection Prevalence : 0.5538
## Balanced Accuracy : 0.5223
##
## 'Positive' Class : spontaneous
##
# model 6A dynamics and eye movement
set.seed(1973)
smile__tree_model_6A <-
train(smile_type ~ onset_mean + apex_mean + offset_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6A$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5522393 0.10505764 0.08946029 0.17418506
## 2 0.06666667 0.5314728 0.06480625 0.09078317 0.17739898
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6A$coefnames
## [1] "onset_mean" "apex_mean" "offset_mean" "eye_mean"
varImp(smile__tree_model_6A)
## rpart variable importance
##
## Overall
## eye_mean 100.00
## offset_mean 71.88
## onset_mean 29.55
## apex_mean 0.00
# summary(smile__tree_model_6A$finalModel)
fancyRpartPlot(smile__tree_model_6A$finalModel)
smile__tree_model_6A_pred <- predict(smile__tree_model_6A, tst_smile)
summary(smile__tree_model_6A_pred)
## spontaneous deliberate
## 88 54
smile__tree_model_6A_confM <- confusionMatrix(
smile__tree_model_6A_pred,
tst_smile$smile_type
)
smile__tree_model_6A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 45 43
## deliberate 25 29
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.40080
##
## Kappa : 0.0455
##
## Mcnemar's Test P-Value : 0.03925
##
## Sensitivity : 0.6429
## Specificity : 0.4028
## Pos Pred Value : 0.5114
## Neg Pred Value : 0.5370
## Prevalence : 0.4930
## Detection Rate : 0.3169
## Detection Prevalence : 0.6197
## Balanced Accuracy : 0.5228
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6A.1_pred <- predict(smile__tree_model_6A, tst_smile_boys)
summary(smile__tree_model_6A.1_pred)
## spontaneous deliberate
## 52 25
smile__tree_model_6A.1_confM <- confusionMatrix(
smile__tree_model_6A.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 26
## deliberate 11 14
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.54593
##
## Kappa : 0.0519
##
## Mcnemar's Test P-Value : 0.02136
##
## Sensitivity : 0.7027
## Specificity : 0.3500
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5600
## Prevalence : 0.4805
## Detection Rate : 0.3377
## Detection Prevalence : 0.6753
## Balanced Accuracy : 0.5264
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6A.2_pred <- predict(smile__tree_model_6A, tst_smile_girls)
summary(smile__tree_model_6A.2_pred)
## spontaneous deliberate
## 36 29
smile__tree_model_6A.2_confM <- confusionMatrix(
smile__tree_model_6A.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 17
## deliberate 14 15
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0446
##
## Mcnemar's Test P-Value : 0.7194
##
## Sensitivity : 0.5758
## Specificity : 0.4688
## Pos Pred Value : 0.5278
## Neg Pred Value : 0.5172
## Prevalence : 0.5077
## Detection Rate : 0.2923
## Detection Prevalence : 0.5538
## Balanced Accuracy : 0.5223
##
## 'Positive' Class : spontaneous
##
# model 6B dynamics and lip movement
set.seed(1973)
smile__tree_model_6B <-
train(smile_type ~ onset_mean + apex_mean + offset_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6B$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5341466 0.06676405 0.08133394 0.1623784
## 2 0.06666667 0.5102718 0.02342038 0.08905266 0.1730390
## 3 0.07878788 0.5106283 0.02562619 0.06235534 0.1200097
smile__tree_model_6B$coefnames
## [1] "onset_mean" "apex_mean" "offset_mean" "lip_mean"
varImp(smile__tree_model_6B)
## rpart variable importance
##
## Overall
## apex_mean 100.00
## offset_mean 67.01
## lip_mean 64.83
## onset_mean 0.00
# summary(smile__tree_model_6B$finalModel)
fancyRpartPlot(smile__tree_model_6B$finalModel)
smile__tree_model_6B_pred <- predict(smile__tree_model_6B, tst_smile)
summary(smile__tree_model_6B_pred)
## spontaneous deliberate
## 51 91
smile__tree_model_6B_confM <- confusionMatrix(
smile__tree_model_6B_pred,
tst_smile$smile_type
)
smile__tree_model_6B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 36 15
## deliberate 34 57
##
## Accuracy : 0.6549
## 95% CI : (0.5706, 0.7326)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0002621
##
## Kappa : 0.3071
##
## Mcnemar's Test P-Value : 0.0101280
##
## Sensitivity : 0.5143
## Specificity : 0.7917
## Pos Pred Value : 0.7059
## Neg Pred Value : 0.6264
## Prevalence : 0.4930
## Detection Rate : 0.2535
## Detection Prevalence : 0.3592
## Balanced Accuracy : 0.6530
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6B.1_pred <- predict(smile__tree_model_6B, tst_smile_boys)
# summary(smile__tree_model_6B.1_pred)
smile__tree_model_6B.1_confM <- confusionMatrix(
smile__tree_model_6B.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 10
## deliberate 17 30
##
## Accuracy : 0.6494
## 95% CI : (0.5322, 0.7547)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.01457
##
## Kappa : 0.2926
##
## Mcnemar's Test P-Value : 0.24821
##
## Sensitivity : 0.5405
## Specificity : 0.7500
## Pos Pred Value : 0.6667
## Neg Pred Value : 0.6383
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.3896
## Balanced Accuracy : 0.6453
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6B.2_pred <- predict(smile__tree_model_6B, tst_smile_girls)
summary(smile__tree_model_6B.2_pred)
## spontaneous deliberate
## 21 44
smile__tree_model_6B.2_confM <- confusionMatrix(
smile__tree_model_6B.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 5
## deliberate 17 27
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3267
##
## Mcnemar's Test P-Value : 0.019016
##
## Sensitivity : 0.4848
## Specificity : 0.8438
## Pos Pred Value : 0.7619
## Neg Pred Value : 0.6136
## Prevalence : 0.5077
## Detection Rate : 0.2462
## Detection Prevalence : 0.3231
## Balanced Accuracy : 0.6643
##
## 'Positive' Class : spontaneous
##
# model 6C onset and movement
set.seed(1973)
smile__tree_model_6C <- train(smile_type ~ onset_mean + eye_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6C$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02424242 0.5678420 0.13387329 0.07670856 0.15409631
## 2 0.06666667 0.5863024 0.16881293 0.10339900 0.20811509
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6C$coefnames
## [1] "onset_mean" "eye_mean" "lip_mean"
varImp(smile__tree_model_6C)
## rpart variable importance
##
## Overall
## eye_mean 100.0
## onset_mean 18.8
## lip_mean 0.0
# summary(smile__tree_model_6C$finalModel)
fancyRpartPlot(smile__tree_model_6C$finalModel)
smile__tree_model_6C_pred <- predict(smile__tree_model_6C, tst_smile)
summary(smile__tree_model_6C_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_6C_confM <- confusionMatrix(
smile__tree_model_6C_pred,
tst_smile$smile_type
)
smile__tree_model_6C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 45 56
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.076642
##
## Kappa : 0.1357
##
## Mcnemar's Test P-Value : 0.000337
##
## Sensitivity : 0.3571
## Specificity : 0.7778
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.5545
## Prevalence : 0.4930
## Detection Rate : 0.1761
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5675
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6C.1_pred <- predict(smile__tree_model_6C, tst_smile_boys)
summary(smile__tree_model_6C.1_pred)
## spontaneous deliberate
## 22 55
smile__tree_model_6C.1_confM <- confusionMatrix(
smile__tree_model_6C.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 9
## deliberate 24 31
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.21259
##
## Kappa : 0.1283
##
## Mcnemar's Test P-Value : 0.01481
##
## Sensitivity : 0.3514
## Specificity : 0.7750
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.5636
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2857
## Balanced Accuracy : 0.5632
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6C.2_pred <- predict(smile__tree_model_6C, tst_smile_girls)
summary(smile__tree_model_6C.2_pred)
## spontaneous deliberate
## 19 46
smile__tree_model_6C.2_confM <- confusionMatrix(
smile__tree_model_6C.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 7
## deliberate 21 25
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.19273
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.01402
##
## Sensitivity : 0.3636
## Specificity : 0.7812
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.5435
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2923
## Balanced Accuracy : 0.5724
##
## 'Positive' Class : spontaneous
##
# model 6D onset + eye
set.seed(1973)
smile__tree_model_6D <- train(smile_type ~ onset_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6D$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5830882 0.16410842 0.07002286 0.13877600
## 2 0.06666667 0.5863024 0.16881293 0.10339900 0.20811509
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6D$coefnames
## [1] "onset_mean" "eye_mean"
varImp(smile__tree_model_6D)
## rpart variable importance
##
## Overall
## eye_mean 100
## onset_mean 0
# summary(smile__tree_model_6D$finalModel)
fancyRpartPlot(smile__tree_model_6D$finalModel)
smile__tree_model_6D_pred <- predict(smile__tree_model_6D, tst_smile)
summary(smile__tree_model_6D_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_6D_confM <- confusionMatrix(
smile__tree_model_6D_pred,
tst_smile$smile_type
)
smile__tree_model_6D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 45 56
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.076642
##
## Kappa : 0.1357
##
## Mcnemar's Test P-Value : 0.000337
##
## Sensitivity : 0.3571
## Specificity : 0.7778
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.5545
## Prevalence : 0.4930
## Detection Rate : 0.1761
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5675
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6D.1_pred <- predict(smile__tree_model_6D, tst_smile_boys)
summary(smile__tree_model_6D.1_pred)
## spontaneous deliberate
## 22 55
smile__tree_model_6D.1_confM <- confusionMatrix(
smile__tree_model_6D.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 9
## deliberate 24 31
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.21259
##
## Kappa : 0.1283
##
## Mcnemar's Test P-Value : 0.01481
##
## Sensitivity : 0.3514
## Specificity : 0.7750
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.5636
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2857
## Balanced Accuracy : 0.5632
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6D.2_pred <- predict(smile__tree_model_6D, tst_smile_girls)
summary(smile__tree_model_6D.2_pred)
## spontaneous deliberate
## 19 46
smile__tree_model_6D.2_confM <- confusionMatrix(
smile__tree_model_6D.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 7
## deliberate 21 25
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.19273
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.01402
##
## Sensitivity : 0.3636
## Specificity : 0.7812
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.5435
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2923
## Balanced Accuracy : 0.5724
##
## 'Positive' Class : spontaneous
##
# model 6E onset + lip
set.seed(1973)
smile__tree_model_6E <- train(smile_type ~ onset_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6E$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02121212 0.5669285 0.13279284 0.05096836 0.1021507
## 2 0.02424242 0.5670232 0.13415123 0.04546484 0.0904795
## 3 0.05909091 0.4897504 -0.02427537 0.05640185 0.1111774
smile__tree_model_6E$coefnames
## [1] "onset_mean" "lip_mean"
varImp(smile__tree_model_6E)
## rpart variable importance
##
## Overall
## lip_mean 100
## onset_mean 0
# summary(smile__tree_model_6E$finalModel)
fancyRpartPlot(smile__tree_model_6E$finalModel)
smile__tree_model_6E_pred <- predict(smile__tree_model_6E, tst_smile)
summary(smile__tree_model_6E_pred)
## spontaneous deliberate
## 68 74
smile__tree_model_6E_confM <- confusionMatrix(
smile__tree_model_6E_pred,
tst_smile$smile_type
)
smile__tree_model_6E_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 35 33
## deliberate 35 39
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.4008
##
## Kappa : 0.0417
##
## Mcnemar's Test P-Value : 0.9035
##
## Sensitivity : 0.5000
## Specificity : 0.5417
## Pos Pred Value : 0.5147
## Neg Pred Value : 0.5270
## Prevalence : 0.4930
## Detection Rate : 0.2465
## Detection Prevalence : 0.4789
## Balanced Accuracy : 0.5208
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6E.1_pred <- predict(smile__tree_model_6E, tst_smile_boys)
summary(smile__tree_model_6E.1_pred)
## spontaneous deliberate
## 34 43
smile__tree_model_6E.1_confM <- confusionMatrix(
smile__tree_model_6E.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 15 19
## deliberate 22 21
##
## Accuracy : 0.4675
## 95% CI : (0.3529, 0.5848)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.8476
##
## Kappa : -0.0698
##
## Mcnemar's Test P-Value : 0.7548
##
## Sensitivity : 0.4054
## Specificity : 0.5250
## Pos Pred Value : 0.4412
## Neg Pred Value : 0.4884
## Prevalence : 0.4805
## Detection Rate : 0.1948
## Detection Prevalence : 0.4416
## Balanced Accuracy : 0.4652
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6E.2_pred <- predict(smile__tree_model_6E, tst_smile_girls)
summary(smile__tree_model_6E.2_pred)
## spontaneous deliberate
## 34 31
smile__tree_model_6E.2_confM <- confusionMatrix(
smile__tree_model_6E.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 14
## deliberate 13 18
##
## Accuracy : 0.5846
## 95% CI : (0.4556, 0.7056)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.132
##
## Kappa : 0.1686
##
## Mcnemar's Test P-Value : 1.000
##
## Sensitivity : 0.6061
## Specificity : 0.5625
## Pos Pred Value : 0.5882
## Neg Pred Value : 0.5806
## Prevalence : 0.5077
## Detection Rate : 0.3077
## Detection Prevalence : 0.5231
## Balanced Accuracy : 0.5843
##
## 'Positive' Class : spontaneous
##
# model 6F apex and movement
set.seed(1973)
smile__tree_model_6F <- train(smile_type ~ apex_mean + eye_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6F$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5920065 0.18186252 0.09150065 0.18186377
## 2 0.06666667 0.5892435 0.17469528 0.10479850 0.21101409
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6F$coefnames
## [1] "apex_mean" "eye_mean" "lip_mean"
varImp(smile__tree_model_6F)
## rpart variable importance
##
## Overall
## eye_mean 100.00
## apex_mean 19.36
## lip_mean 0.00
# summary(smile__tree_model_6F$finalModel)
fancyRpartPlot(smile__tree_model_6F$finalModel,
caption = "model 6F: Decision Tree - apex, eye, lip"
)
smile__tree_model_6F_pred <- predict(smile__tree_model_6F, tst_smile)
summary(smile__tree_model_6F_pred)
## spontaneous deliberate
## 88 54
smile__tree_model_6F_confM <- confusionMatrix(
smile__tree_model_6F_pred,
tst_smile$smile_type
)
smile__tree_model_6F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 45 43
## deliberate 25 29
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.40080
##
## Kappa : 0.0455
##
## Mcnemar's Test P-Value : 0.03925
##
## Sensitivity : 0.6429
## Specificity : 0.4028
## Pos Pred Value : 0.5114
## Neg Pred Value : 0.5370
## Prevalence : 0.4930
## Detection Rate : 0.3169
## Detection Prevalence : 0.6197
## Balanced Accuracy : 0.5228
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6F.1_pred <- predict(smile__tree_model_6F, tst_smile_boys)
summary(smile__tree_model_6F.1_pred)
## spontaneous deliberate
## 52 25
smile__tree_model_6F.1_confM <- confusionMatrix(
smile__tree_model_6F.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 26
## deliberate 11 14
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.54593
##
## Kappa : 0.0519
##
## Mcnemar's Test P-Value : 0.02136
##
## Sensitivity : 0.7027
## Specificity : 0.3500
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5600
## Prevalence : 0.4805
## Detection Rate : 0.3377
## Detection Prevalence : 0.6753
## Balanced Accuracy : 0.5264
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6F.2_pred <- predict(smile__tree_model_6F, tst_smile_girls)
summary(smile__tree_model_6F.2_pred)
## spontaneous deliberate
## 36 29
smile__tree_model_6F.2_confM <- confusionMatrix(
smile__tree_model_6F.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 17
## deliberate 14 15
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0446
##
## Mcnemar's Test P-Value : 0.7194
##
## Sensitivity : 0.5758
## Specificity : 0.4688
## Pos Pred Value : 0.5278
## Neg Pred Value : 0.5172
## Prevalence : 0.5077
## Detection Rate : 0.2923
## Detection Prevalence : 0.5538
## Balanced Accuracy : 0.5223
##
## 'Positive' Class : spontaneous
##
# model 6G apex + eye
set.seed(1973)
smile__tree_model_6G <- train(smile_type ~ apex_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6G$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5920065 0.18186252 0.09150065 0.18186377
## 2 0.06666667 0.5892435 0.17469528 0.10479850 0.21101409
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6G$coefnames
## [1] "apex_mean" "eye_mean"
varImp(smile__tree_model_6G)
## rpart variable importance
##
## Overall
## eye_mean 100
## apex_mean 0
# summary(smile__tree_model_6G$finalModel)
fancyRpartPlot(smile__tree_model_6G$finalModel)
smile__tree_model_6G_pred <- predict(smile__tree_model_6G, tst_smile)
summary(smile__tree_model_6G_pred)
## spontaneous deliberate
## 88 54
smile__tree_model_6G_confM <- confusionMatrix(
smile__tree_model_6G_pred,
tst_smile$smile_type
)
smile__tree_model_6G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 45 43
## deliberate 25 29
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.40080
##
## Kappa : 0.0455
##
## Mcnemar's Test P-Value : 0.03925
##
## Sensitivity : 0.6429
## Specificity : 0.4028
## Pos Pred Value : 0.5114
## Neg Pred Value : 0.5370
## Prevalence : 0.4930
## Detection Rate : 0.3169
## Detection Prevalence : 0.6197
## Balanced Accuracy : 0.5228
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6G.1_pred <- predict(smile__tree_model_6G, tst_smile_boys)
summary(smile__tree_model_6G.1_pred)
## spontaneous deliberate
## 52 25
smile__tree_model_6G.1_confM <- confusionMatrix(
smile__tree_model_6G.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 26
## deliberate 11 14
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.54593
##
## Kappa : 0.0519
##
## Mcnemar's Test P-Value : 0.02136
##
## Sensitivity : 0.7027
## Specificity : 0.3500
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5600
## Prevalence : 0.4805
## Detection Rate : 0.3377
## Detection Prevalence : 0.6753
## Balanced Accuracy : 0.5264
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6G.2_pred <- predict(smile__tree_model_6G, tst_smile_girls)
summary(smile__tree_model_6G.2_pred)
## spontaneous deliberate
## 36 29
smile__tree_model_6G.2_confM <- confusionMatrix(
smile__tree_model_6G.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 17
## deliberate 14 15
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0446
##
## Mcnemar's Test P-Value : 0.7194
##
## Sensitivity : 0.5758
## Specificity : 0.4688
## Pos Pred Value : 0.5278
## Neg Pred Value : 0.5172
## Prevalence : 0.5077
## Detection Rate : 0.2923
## Detection Prevalence : 0.5538
## Balanced Accuracy : 0.5223
##
## 'Positive' Class : spontaneous
##
# model 6H apex + lip
set.seed(1973)
smile__tree_model_6H <- train(smile_type ~ apex_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6H$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02272727 0.4543895 -0.09300652 0.09708536 0.1932817
## 2 0.03333333 0.4632074 -0.07541557 0.09095422 0.1816019
## 3 0.11515152 0.4866310 -0.03605245 0.03749912 0.0690707
smile__tree_model_6H$coefnames
## [1] "apex_mean" "lip_mean"
varImp(smile__tree_model_6H)
## rpart variable importance
##
## Overall
## apex_mean NaN
## lip_mean NaN
# summary(smile__tree_model_6H$finalModel)
smile__tree_model_6H_pred <- predict(smile__tree_model_6H, tst_smile)
summary(smile__tree_model_6H_pred)
## spontaneous deliberate
## 0 142
smile__tree_model_6H_confM <- confusionMatrix(
smile__tree_model_6H_pred,
tst_smile$smile_type
)
smile__tree_model_6H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 70 72
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.000
## Specificity : 1.000
## Pos Pred Value : NaN
## Neg Pred Value : 0.507
## Prevalence : 0.493
## Detection Rate : 0.000
## Detection Prevalence : 0.000
## Balanced Accuracy : 0.500
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6H.1_pred <- predict(smile__tree_model_6H, tst_smile_boys)
summary(smile__tree_model_6H.1_pred)
## spontaneous deliberate
## 0 77
smile__tree_model_6H.1_confM <- confusionMatrix(
smile__tree_model_6H.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 37 40
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 3.252e-09
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.5195
## Prevalence : 0.4805
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6H.2_pred <- predict(smile__tree_model_6H, tst_smile_girls)
summary(smile__tree_model_6H.2_pred)
## spontaneous deliberate
## 0 65
smile__tree_model_6H.2_confM <- confusionMatrix(
smile__tree_model_6H.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 0 0
## deliberate 33 32
##
## Accuracy : 0.4923
## 95% CI : (0.366, 0.6193)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.6452
##
## Kappa : 0
##
## Mcnemar's Test P-Value : 2.54e-08
##
## Sensitivity : 0.0000
## Specificity : 1.0000
## Pos Pred Value : NaN
## Neg Pred Value : 0.4923
## Prevalence : 0.5077
## Detection Rate : 0.0000
## Detection Prevalence : 0.0000
## Balanced Accuracy : 0.5000
##
## 'Positive' Class : spontaneous
##
# model 6I offset and movement
set.seed(1973)
smile__tree_model_6I <- train(smile_type ~ offset_mean + eye_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6I$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02272727 0.5322081 0.06456196 0.09174672 0.18428052
## 2 0.06666667 0.5407587 0.08370581 0.10511016 0.20533476
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6I$coefnames
## [1] "offset_mean" "eye_mean" "lip_mean"
varImp(smile__tree_model_6I)
## rpart variable importance
##
## Overall
## eye_mean 100.00
## offset_mean 66.59
## lip_mean 0.00
# summary(smile__tree_model_6I$finalModel)
fancyRpartPlot(smile__tree_model_6I$finalModel)
smile__tree_model_6I_pred <- predict(smile__tree_model_6I, tst_smile)
summary(smile__tree_model_6I_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_6I_confM <- confusionMatrix(
smile__tree_model_6I_pred,
tst_smile$smile_type
)
smile__tree_model_6I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 45 56
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.076642
##
## Kappa : 0.1357
##
## Mcnemar's Test P-Value : 0.000337
##
## Sensitivity : 0.3571
## Specificity : 0.7778
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.5545
## Prevalence : 0.4930
## Detection Rate : 0.1761
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5675
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6I.1_pred <- predict(smile__tree_model_6I, tst_smile_boys)
summary(smile__tree_model_6I.1_pred)
## spontaneous deliberate
## 22 55
smile__tree_model_6I.1_confM <- confusionMatrix(
smile__tree_model_6I.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 9
## deliberate 24 31
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.21259
##
## Kappa : 0.1283
##
## Mcnemar's Test P-Value : 0.01481
##
## Sensitivity : 0.3514
## Specificity : 0.7750
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.5636
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2857
## Balanced Accuracy : 0.5632
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6I.2_pred <- predict(smile__tree_model_6I, tst_smile_girls)
summary(smile__tree_model_6I.2_pred)
## spontaneous deliberate
## 19 46
smile__tree_model_6I.2_confM <- confusionMatrix(
smile__tree_model_6I.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 7
## deliberate 21 25
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.19273
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.01402
##
## Sensitivity : 0.3636
## Specificity : 0.7812
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.5435
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2923
## Balanced Accuracy : 0.5724
##
## 'Positive' Class : spontaneous
##
# model 6J offset + eye
set.seed(1973)
smile__tree_model_6J <- train(smile_type ~ offset_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6J$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5642547 0.12621182 0.07907261 0.15669999
## 2 0.06666667 0.5862132 0.16762689 0.10072271 0.20239814
## 3 0.16363636 0.4868093 -0.03388104 0.03960046 0.07404332
smile__tree_model_6J$coefnames
## [1] "offset_mean" "eye_mean"
varImp(smile__tree_model_6J)
## rpart variable importance
##
## Overall
## eye_mean 100
## offset_mean 0
# summary(smile__tree_model_6J$finalModel)
fancyRpartPlot(smile__tree_model_6J$finalModel)
smile__tree_model_6J_pred <- predict(smile__tree_model_6J, tst_smile)
summary(smile__tree_model_6J_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_6J_confM <- confusionMatrix(
smile__tree_model_6J_pred,
tst_smile$smile_type
)
smile__tree_model_6J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 45 56
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.076642
##
## Kappa : 0.1357
##
## Mcnemar's Test P-Value : 0.000337
##
## Sensitivity : 0.3571
## Specificity : 0.7778
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.5545
## Prevalence : 0.4930
## Detection Rate : 0.1761
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5675
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6J.1_pred <- predict(smile__tree_model_6J, tst_smile_boys)
summary(smile__tree_model_6J.1_pred)
## spontaneous deliberate
## 22 55
smile__tree_model_6J.1_confM <- confusionMatrix(
smile__tree_model_6J.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 9
## deliberate 24 31
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.21259
##
## Kappa : 0.1283
##
## Mcnemar's Test P-Value : 0.01481
##
## Sensitivity : 0.3514
## Specificity : 0.7750
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.5636
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2857
## Balanced Accuracy : 0.5632
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6J.2_pred <- predict(smile__tree_model_6J, tst_smile_girls)
summary(smile__tree_model_6J.2_pred)
## spontaneous deliberate
## 19 46
smile__tree_model_6J.2_confM <- confusionMatrix(
smile__tree_model_6J.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 7
## deliberate 21 25
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.19273
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.01402
##
## Sensitivity : 0.3636
## Specificity : 0.7812
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.5435
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2923
## Balanced Accuracy : 0.5724
##
## 'Positive' Class : spontaneous
##
# model 6K offset + lip
set.seed(1973)
smile__tree_model_6K <- train(smile_type ~ offset_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_6K$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.02424242 0.5676693 0.13426752 0.07837002 0.1574382
## 2 0.03434343 0.5552640 0.11098029 0.08076035 0.1634492
## 3 0.07878788 0.5075869 0.02036637 0.08687993 0.1683719
smile__tree_model_6K$coefnames
## [1] "offset_mean" "lip_mean"
varImp(smile__tree_model_6K)
## rpart variable importance
##
## Overall
## lip_mean 100
## offset_mean 0
# summary(smile__tree_model_6K$finalModel)
fancyRpartPlot(smile__tree_model_6K$finalModel)
smile__tree_model_6K_pred <- predict(smile__tree_model_6K, tst_smile)
summary(smile__tree_model_6K_pred)
## spontaneous deliberate
## 60 82
smile__tree_model_6K_confM <- confusionMatrix(
smile__tree_model_6K_pred,
tst_smile$smile_type
)
smile__tree_model_6K_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 37 23
## deliberate 33 49
##
## Accuracy : 0.6056
## 95% CI : (0.5202, 0.6865)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.01151
##
## Kappa : 0.2095
##
## Mcnemar's Test P-Value : 0.22910
##
## Sensitivity : 0.5286
## Specificity : 0.6806
## Pos Pred Value : 0.6167
## Neg Pred Value : 0.5976
## Prevalence : 0.4930
## Detection Rate : 0.2606
## Detection Prevalence : 0.4225
## Balanced Accuracy : 0.6046
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_6K.1_pred <- predict(smile__tree_model_6K, tst_smile_boys)
summary(smile__tree_model_6K.1_pred)
## spontaneous deliberate
## 35 42
smile__tree_model_6K.1_confM <- confusionMatrix(
smile__tree_model_6K.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_6K.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 15
## deliberate 17 25
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1659
##
## Mcnemar's Test P-Value : 0.8597
##
## Sensitivity : 0.5405
## Specificity : 0.6250
## Pos Pred Value : 0.5714
## Neg Pred Value : 0.5952
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.4545
## Balanced Accuracy : 0.5828
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_6K.2_pred <- predict(smile__tree_model_6K, tst_smile_girls)
summary(smile__tree_model_6K.2_pred)
## spontaneous deliberate
## 25 40
smile__tree_model_6K.2_confM <- confusionMatrix(
smile__tree_model_6K.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_6K.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 17 8
## deliberate 16 24
##
## Accuracy : 0.6308
## 95% CI : (0.502, 0.7472)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.03086
##
## Kappa : 0.2642
##
## Mcnemar's Test P-Value : 0.15304
##
## Sensitivity : 0.5152
## Specificity : 0.7500
## Pos Pred Value : 0.6800
## Neg Pred Value : 0.6000
## Prevalence : 0.5077
## Detection Rate : 0.2615
## Detection Prevalence : 0.3846
## Balanced Accuracy : 0.6326
##
## 'Positive' Class : spontaneous
##
# Model 7: dynamics and AU's
set.seed(1973)
smile__tree_model_7 <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04848485 0.5368761 0.07420769 0.07460008 0.1488441
## 2 0.05252525 0.5308155 0.05932530 0.07759515 0.1551832
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7$coefnames
## [1] "AU01_r_mean" "AU02_r_mean" "AU04_r_mean" "AU05_r_mean" "AU06_r_mean"
## [6] "AU07_r_mean" "AU09_r_mean" "AU10_r_mean" "AU12_r_mean" "AU14_r_mean"
## [11] "AU15_r_mean" "AU17_r_mean" "AU20_r_mean" "AU23_r_mean" "AU25_r_mean"
## [16] "AU26_r_mean" "AU45_r_mean" "onset_mean" "apex_mean" "offset_mean"
varImp(smile__tree_model_7)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 81.66
## AU25_r_mean 63.30
## AU09_r_mean 56.35
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## apex_mean 38.50
## AU14_r_mean 21.58
## AU12_r_mean 0.00
## AU23_r_mean 0.00
## AU07_r_mean 0.00
## AU02_r_mean 0.00
## AU04_r_mean 0.00
## offset_mean 0.00
## AU06_r_mean 0.00
## onset_mean 0.00
## AU17_r_mean 0.00
## AU20_r_mean 0.00
## AU26_r_mean 0.00
## AU15_r_mean 0.00
# summary(smile__tree_model_7$finalModel)
fancyRpartPlot(smile__tree_model_7$finalModel)
smile__tree_model_7_pred <- predict(smile__tree_model_7, tst_smile)
summary(smile__tree_model_7_pred)
## spontaneous deliberate
## 84 58
smile__tree_model_7_confM <- confusionMatrix(
smile__tree_model_7_pred,
tst_smile$smile_type
)
smile__tree_model_7_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 40
## deliberate 26 32
##
## Accuracy : 0.5352
## 95% CI : (0.4497, 0.6193)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2786
##
## Kappa : 0.0728
##
## Mcnemar's Test P-Value : 0.1096
##
## Sensitivity : 0.6286
## Specificity : 0.4444
## Pos Pred Value : 0.5238
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5915
## Balanced Accuracy : 0.5365
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7.1_pred <- predict(smile__tree_model_7, tst_smile_boys)
summary(smile__tree_model_7.1_pred)
## spontaneous deliberate
## 40 37
smile__tree_model_7.1_confM <- confusionMatrix(
smile__tree_model_7.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 20
## deliberate 17 20
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0.0404
##
## Mcnemar's Test P-Value : 0.7423
##
## Sensitivity : 0.5405
## Specificity : 0.5000
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5405
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.5195
## Balanced Accuracy : 0.5203
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7.2_pred <- predict(smile__tree_model_7, tst_smile_girls)
summary(smile__tree_model_7.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_7.2_confM <- confusionMatrix(
smile__tree_model_7.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 9 12
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.06332
##
## Sensitivity : 0.7273
## Specificity : 0.3750
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.5511
##
## 'Positive' Class : spontaneous
##
# 7A AU's + onset
set.seed(1973)
smile__tree_model_7A <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
onset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7A$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5636141 0.12907456 0.08437636 0.1685849
## 2 0.05252525 0.5308155 0.05932530 0.07759515 0.1551832
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7A$coefnames
## [1] "AU01_r_mean" "AU02_r_mean" "AU04_r_mean" "AU05_r_mean" "AU06_r_mean"
## [6] "AU07_r_mean" "AU09_r_mean" "AU10_r_mean" "AU12_r_mean" "AU14_r_mean"
## [11] "AU15_r_mean" "AU17_r_mean" "AU20_r_mean" "AU23_r_mean" "AU25_r_mean"
## [16] "AU26_r_mean" "AU45_r_mean" "onset_mean"
varImp(smile__tree_model_7A)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 81.66
## AU25_r_mean 77.78
## AU09_r_mean 73.78
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## AU14_r_mean 21.58
## AU12_r_mean 0.00
## AU23_r_mean 0.00
## AU07_r_mean 0.00
## AU02_r_mean 0.00
## AU04_r_mean 0.00
## AU06_r_mean 0.00
## onset_mean 0.00
## AU17_r_mean 0.00
## AU20_r_mean 0.00
## AU26_r_mean 0.00
## AU15_r_mean 0.00
# summary(smile__tree_model_7A$finalModel)
fancyRpartPlot(smile__tree_model_7A$finalModel)
smile__tree_model_7A_pred <- predict(smile__tree_model_7A, tst_smile)
summary(smile__tree_model_7A_pred)
## spontaneous deliberate
## 84 58
smile__tree_model_7A_confM <- confusionMatrix(
smile__tree_model_7A_pred,
tst_smile$smile_type
)
smile__tree_model_7A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 40
## deliberate 26 32
##
## Accuracy : 0.5352
## 95% CI : (0.4497, 0.6193)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2786
##
## Kappa : 0.0728
##
## Mcnemar's Test P-Value : 0.1096
##
## Sensitivity : 0.6286
## Specificity : 0.4444
## Pos Pred Value : 0.5238
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5915
## Balanced Accuracy : 0.5365
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7A.1_pred <- predict(smile__tree_model_7A, tst_smile_boys)
summary(smile__tree_model_7A.1_pred)
## spontaneous deliberate
## 40 37
smile__tree_model_7A.1_confM <- confusionMatrix(
smile__tree_model_7A.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 20
## deliberate 17 20
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0.0404
##
## Mcnemar's Test P-Value : 0.7423
##
## Sensitivity : 0.5405
## Specificity : 0.5000
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5405
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.5195
## Balanced Accuracy : 0.5203
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7A.2_pred <- predict(smile__tree_model_7A, tst_smile_girls)
summary(smile__tree_model_7A.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_7A.2_confM <- confusionMatrix(
smile__tree_model_7A.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 9 12
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.06332
##
## Sensitivity : 0.7273
## Specificity : 0.3750
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.5511
##
## 'Positive' Class : spontaneous
##
# 7B AU's + apex
set.seed(1973)
smile__tree_model_7B <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
apex_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7B$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04848485 0.5368761 0.07420769 0.07460008 0.1488441
## 2 0.05252525 0.5308155 0.05932530 0.07759515 0.1551832
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7B$coefnames
## [1] "AU01_r_mean" "AU02_r_mean" "AU04_r_mean" "AU05_r_mean" "AU06_r_mean"
## [6] "AU07_r_mean" "AU09_r_mean" "AU10_r_mean" "AU12_r_mean" "AU14_r_mean"
## [11] "AU15_r_mean" "AU17_r_mean" "AU20_r_mean" "AU23_r_mean" "AU25_r_mean"
## [16] "AU26_r_mean" "AU45_r_mean" "apex_mean"
varImp(smile__tree_model_7B)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 81.66
## AU25_r_mean 63.30
## AU09_r_mean 56.35
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## apex_mean 38.50
## AU14_r_mean 21.58
## AU12_r_mean 0.00
## AU23_r_mean 0.00
## AU07_r_mean 0.00
## AU02_r_mean 0.00
## AU04_r_mean 0.00
## AU06_r_mean 0.00
## AU17_r_mean 0.00
## AU20_r_mean 0.00
## AU26_r_mean 0.00
## AU15_r_mean 0.00
# summary(smile__tree_model_7B$finalModel)
fancyRpartPlot(smile__tree_model_7B$finalModel)
smile__tree_model_7B_pred <- predict(smile__tree_model_7B, tst_smile)
summary(smile__tree_model_7B_pred)
## spontaneous deliberate
## 84 58
smile__tree_model_7B_confM <- confusionMatrix(
smile__tree_model_7B_pred,
tst_smile$smile_type
)
smile__tree_model_7B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 40
## deliberate 26 32
##
## Accuracy : 0.5352
## 95% CI : (0.4497, 0.6193)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2786
##
## Kappa : 0.0728
##
## Mcnemar's Test P-Value : 0.1096
##
## Sensitivity : 0.6286
## Specificity : 0.4444
## Pos Pred Value : 0.5238
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5915
## Balanced Accuracy : 0.5365
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7B.1_pred <- predict(smile__tree_model_7B, tst_smile_boys)
summary(smile__tree_model_7B.1_pred)
## spontaneous deliberate
## 40 37
smile__tree_model_7B.1_confM <- confusionMatrix(
smile__tree_model_7B.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 20
## deliberate 17 20
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0.0404
##
## Mcnemar's Test P-Value : 0.7423
##
## Sensitivity : 0.5405
## Specificity : 0.5000
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5405
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.5195
## Balanced Accuracy : 0.5203
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7B.2_pred <- predict(smile__tree_model_7B, tst_smile_girls)
summary(smile__tree_model_7B.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_7B.2_confM <- confusionMatrix(
smile__tree_model_7B.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 9 12
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.06332
##
## Sensitivity : 0.7273
## Specificity : 0.3750
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.5511
##
## 'Positive' Class : spontaneous
##
# 7C AU's + offset
set.seed(1973)
smile__tree_model_7C <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
offset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7C$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5666444 0.13557324 0.08539890 0.1707513
## 2 0.05252525 0.5308155 0.05932530 0.07759515 0.1551832
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7C$coefnames
## [1] "AU01_r_mean" "AU02_r_mean" "AU04_r_mean" "AU05_r_mean" "AU06_r_mean"
## [6] "AU07_r_mean" "AU09_r_mean" "AU10_r_mean" "AU12_r_mean" "AU14_r_mean"
## [11] "AU15_r_mean" "AU17_r_mean" "AU20_r_mean" "AU23_r_mean" "AU25_r_mean"
## [16] "AU26_r_mean" "AU45_r_mean" "offset_mean"
varImp(smile__tree_model_7C)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU01_r_mean 81.66
## AU25_r_mean 77.78
## AU09_r_mean 73.78
## AU10_r_mean 54.54
## AU05_r_mean 52.28
## AU14_r_mean 21.58
## AU12_r_mean 0.00
## AU23_r_mean 0.00
## AU07_r_mean 0.00
## AU02_r_mean 0.00
## AU04_r_mean 0.00
## AU06_r_mean 0.00
## offset_mean 0.00
## AU17_r_mean 0.00
## AU20_r_mean 0.00
## AU26_r_mean 0.00
## AU15_r_mean 0.00
# summary(smile__tree_model_7C$finalModel)
fancyRpartPlot(smile__tree_model_7C$finalModel)
smile__tree_model_7C_pred <- predict(smile__tree_model_7C, tst_smile)
summary(smile__tree_model_7C_pred)
## spontaneous deliberate
## 84 58
smile__tree_model_7C_confM <- confusionMatrix(
smile__tree_model_7C_pred,
tst_smile$smile_type
)
smile__tree_model_7C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 40
## deliberate 26 32
##
## Accuracy : 0.5352
## 95% CI : (0.4497, 0.6193)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2786
##
## Kappa : 0.0728
##
## Mcnemar's Test P-Value : 0.1096
##
## Sensitivity : 0.6286
## Specificity : 0.4444
## Pos Pred Value : 0.5238
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5915
## Balanced Accuracy : 0.5365
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7C.1_pred <- predict(smile__tree_model_7C, tst_smile_boys)
summary(smile__tree_model_7C.1_pred)
## spontaneous deliberate
## 40 37
smile__tree_model_7C.1_confM <- confusionMatrix(
smile__tree_model_7C.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 20
## deliberate 17 20
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.5459
##
## Kappa : 0.0404
##
## Mcnemar's Test P-Value : 0.7423
##
## Sensitivity : 0.5405
## Specificity : 0.5000
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5405
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.5195
## Balanced Accuracy : 0.5203
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7C.2_pred <- predict(smile__tree_model_7C, tst_smile_girls)
summary(smile__tree_model_7C.2_pred)
## spontaneous deliberate
## 44 21
smile__tree_model_7C.2_confM <- confusionMatrix(
smile__tree_model_7C.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 9 12
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1028
##
## Mcnemar's Test P-Value : 0.06332
##
## Sensitivity : 0.7273
## Specificity : 0.3750
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.5511
##
## 'Positive' Class : spontaneous
##
# 7D
set.seed(1973)
smile__tree_model_7D <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7D$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5584837 0.11790741 0.03491922 0.06900019
## 2 0.05454545 0.5493928 0.09982095 0.04575070 0.09183336
## 3 0.18787879 0.5223708 0.04901466 0.03672728 0.07788091
smile__tree_model_7D$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "onset_mean" "apex_mean"
## [6] "offset_mean"
varImp(smile__tree_model_7D)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## offset_mean 69.75
## AU06_r_mean 48.99
## onset_mean 46.87
## apex_mean 39.60
## AU12_r_mean 0.00
# summary(smile__tree_model_7D$finalModel)
fancyRpartPlot(smile__tree_model_7D$finalModel)
smile__tree_model_7D_pred <- predict(smile__tree_model_7D, tst_smile)
summary(smile__tree_model_7D_pred)
## spontaneous deliberate
## 92 50
smile__tree_model_7D_confM <- confusionMatrix(
smile__tree_model_7D_pred,
tst_smile$smile_type
)
smile__tree_model_7D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 46
## deliberate 24 26
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.53358
##
## Kappa : 0.0182
##
## Mcnemar's Test P-Value : 0.01207
##
## Sensitivity : 0.6571
## Specificity : 0.3611
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5200
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.6479
## Balanced Accuracy : 0.5091
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7D.1_pred <- predict(smile__tree_model_7D, tst_smile_boys)
summary(smile__tree_model_7D.1_pred)
## spontaneous deliberate
## 49 28
smile__tree_model_7D.1_confM <- confusionMatrix(
smile__tree_model_7D.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 25
## deliberate 13 15
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.63425
##
## Kappa : 0.0234
##
## Mcnemar's Test P-Value : 0.07435
##
## Sensitivity : 0.6486
## Specificity : 0.3750
## Pos Pred Value : 0.4898
## Neg Pred Value : 0.5357
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.6364
## Balanced Accuracy : 0.5118
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7D.2_pred <- predict(smile__tree_model_7D, tst_smile_girls)
summary(smile__tree_model_7D.2_pred)
## spontaneous deliberate
## 43 22
smile__tree_model_7D.2_confM <- confusionMatrix(
smile__tree_model_7D.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 21
## deliberate 11 11
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0105
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.6667
## Specificity : 0.3438
## Pos Pred Value : 0.5116
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6615
## Balanced Accuracy : 0.5052
##
## 'Positive' Class : spontaneous
##
# 7E
set.seed(1973)
smile__tree_model_7E <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7E$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5707832 0.14319007 0.03027685 0.06296834
## 2 0.06060606 0.5464516 0.09717643 0.04388348 0.08896946
## 3 0.18787879 0.5223708 0.04901466 0.03672728 0.07788091
smile__tree_model_7E$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "onset_mean"
varImp(smile__tree_model_7E)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU06_r_mean 31.92
## onset_mean 29.09
## AU12_r_mean 0.00
# summary(smile__tree_model_7E$finalModel)
fancyRpartPlot(smile__tree_model_7E$finalModel)
smile__tree_model_7E_pred <- predict(smile__tree_model_7E, tst_smile)
summary(smile__tree_model_7E_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_7E_confM <- confusionMatrix(
smile__tree_model_7E_pred,
tst_smile$smile_type
)
smile__tree_model_7E_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7E.1_pred <- predict(smile__tree_model_7E, tst_smile_boys)
summary(smile__tree_model_7E.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_7E.1_confM <- confusionMatrix(
smile__tree_model_7E.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7E.2_pred <- predict(smile__tree_model_7E, tst_smile_girls)
summary(smile__tree_model_7E.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_7E.2_confM <- confusionMatrix(
smile__tree_model_7E.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# 7F
set.seed(1973)
smile__tree_model_7F <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + apex_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7F$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5706941 0.14076293 0.03491488 0.07078494
## 2 0.06060606 0.5464516 0.09717643 0.04388348 0.08896946
## 3 0.18787879 0.5223708 0.04901466 0.03672728 0.07788091
smile__tree_model_7F$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "apex_mean"
varImp(smile__tree_model_7F)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## AU06_r_mean 31.92
## apex_mean 19.39
## AU12_r_mean 0.00
# summary(smile__tree_model_7F$finalModel)
fancyRpartPlot(smile__tree_model_7F$finalModel)
smile__tree_model_7F_pred <- predict(smile__tree_model_7F, tst_smile)
summary(smile__tree_model_7F_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_7F_confM <- confusionMatrix(
smile__tree_model_7F_pred,
tst_smile$smile_type
)
smile__tree_model_7F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7F.1_pred <- predict(smile__tree_model_7F, tst_smile_boys)
summary(smile__tree_model_7F.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_7F.1_confM <- confusionMatrix(
smile__tree_model_7F.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7F.2_pred <- predict(smile__tree_model_7F, tst_smile_girls)
summary(smile__tree_model_7F.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_7F.2_confM <- confusionMatrix(
smile__tree_model_7F.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# 7G
set.seed(1973)
smile__tree_model_7G <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + offset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7G$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04848485 0.5524231 0.10692105 0.04645894 0.09414045
## 2 0.05454545 0.5493928 0.10144561 0.04575070 0.09188355
## 3 0.18787879 0.5223708 0.04901466 0.03672728 0.07788091
smile__tree_model_7G$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "offset_mean"
varImp(smile__tree_model_7G)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## offset_mean 59.63
## AU06_r_mean 31.92
## AU12_r_mean 0.00
# summary(smile__tree_model_7G$finalModel)
fancyRpartPlot(smile__tree_model_7G$finalModel)
smile__tree_model_7G_pred <- predict(smile__tree_model_7G, tst_smile)
summary(smile__tree_model_7G_pred)
## spontaneous deliberate
## 92 50
smile__tree_model_7G_confM <- confusionMatrix(
smile__tree_model_7G_pred,
tst_smile$smile_type
)
smile__tree_model_7G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 46
## deliberate 24 26
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.53358
##
## Kappa : 0.0182
##
## Mcnemar's Test P-Value : 0.01207
##
## Sensitivity : 0.6571
## Specificity : 0.3611
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5200
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.6479
## Balanced Accuracy : 0.5091
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7G.1_pred <- predict(smile__tree_model_7G, tst_smile_boys)
summary(smile__tree_model_7G.1_pred)
## spontaneous deliberate
## 49 28
smile__tree_model_7G.1_confM <- confusionMatrix(
smile__tree_model_7G.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 25
## deliberate 13 15
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.63425
##
## Kappa : 0.0234
##
## Mcnemar's Test P-Value : 0.07435
##
## Sensitivity : 0.6486
## Specificity : 0.3750
## Pos Pred Value : 0.4898
## Neg Pred Value : 0.5357
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.6364
## Balanced Accuracy : 0.5118
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7G.2_pred <- predict(smile__tree_model_7G, tst_smile_girls)
summary(smile__tree_model_7G.2_pred)
## spontaneous deliberate
## 43 22
smile__tree_model_7G.2_confM <- confusionMatrix(
smile__tree_model_7G.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 21
## deliberate 11 11
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0105
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.6667
## Specificity : 0.3438
## Pos Pred Value : 0.5116
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6615
## Balanced Accuracy : 0.5052
##
## 'Positive' Class : spontaneous
##
# 7H dynamics and AU's selection
set.seed(1973)
smile__tree_model_7H <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7H$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5425802 0.08849963 0.07231989 0.1450692
## 2 0.04545455 0.5424020 0.08399104 0.07351021 0.1485909
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7H$coefnames
## [1] "AU01_r_mean" "AU09_r_mean" "AU10_r_mean" "AU25_r_mean" "AU45_r_mean"
## [6] "onset_mean" "apex_mean" "offset_mean"
varImp(smile__tree_model_7H)
## rpart variable importance
##
## Overall
## AU25_r_mean 100.00
## AU45_r_mean 90.39
## AU01_r_mean 85.32
## AU09_r_mean 72.27
## apex_mean 55.42
## AU10_r_mean 38.91
## onset_mean 0.00
## offset_mean 0.00
# summary(smile__tree_model_7H$finalModel)
fancyRpartPlot(smile__tree_model_7H$finalModel)
smile__tree_model_7H_pred <- predict(smile__tree_model_7H, tst_smile)
summary(smile__tree_model_7H_pred)
## spontaneous deliberate
## 88 54
smile__tree_model_7H_confM <- confusionMatrix(
smile__tree_model_7H_pred,
tst_smile$smile_type
)
smile__tree_model_7H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 49 39
## deliberate 21 33
##
## Accuracy : 0.5775
## 95% CI : (0.4918, 0.6598)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.05517
##
## Kappa : 0.1578
##
## Mcnemar's Test P-Value : 0.02819
##
## Sensitivity : 0.7000
## Specificity : 0.4583
## Pos Pred Value : 0.5568
## Neg Pred Value : 0.6111
## Prevalence : 0.4930
## Detection Rate : 0.3451
## Detection Prevalence : 0.6197
## Balanced Accuracy : 0.5792
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7H.1_pred <- predict(smile__tree_model_7H, tst_smile_boys)
summary(smile__tree_model_7H.1_pred)
## spontaneous deliberate
## 39 38
smile__tree_model_7H.1_confM <- confusionMatrix(
smile__tree_model_7H.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 17
## deliberate 15 23
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1693
##
## Mcnemar's Test P-Value : 0.8597
##
## Sensitivity : 0.5946
## Specificity : 0.5750
## Pos Pred Value : 0.5641
## Neg Pred Value : 0.6053
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.5065
## Balanced Accuracy : 0.5848
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7H.2_pred <- predict(smile__tree_model_7H, tst_smile_girls)
summary(smile__tree_model_7H.2_pred)
## spontaneous deliberate
## 49 16
smile__tree_model_7H.2_confM <- confusionMatrix(
smile__tree_model_7H.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 27 22
## deliberate 6 10
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.192728
##
## Kappa : 0.1317
##
## Mcnemar's Test P-Value : 0.004586
##
## Sensitivity : 0.8182
## Specificity : 0.3125
## Pos Pred Value : 0.5510
## Neg Pred Value : 0.6250
## Prevalence : 0.5077
## Detection Rate : 0.4154
## Detection Prevalence : 0.7538
## Balanced Accuracy : 0.5653
##
## 'Positive' Class : spontaneous
##
# 7I
set.seed(1973)
smile__tree_model_7I <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7I$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5458890 0.09560258 0.06427069 0.1274249
## 2 0.04545455 0.5364305 0.07335263 0.07508962 0.1507522
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7I$coefnames
## [1] "AU01_r_mean" "AU09_r_mean" "AU10_r_mean" "AU25_r_mean" "AU45_r_mean"
## [6] "onset_mean"
varImp(smile__tree_model_7I)
## rpart variable importance
##
## Overall
## AU25_r_mean 100.00
## AU45_r_mean 90.39
## AU01_r_mean 85.32
## AU10_r_mean 76.48
## AU09_r_mean 72.27
## onset_mean 0.00
# summary(smile__tree_model_7I$finalModel)
fancyRpartPlot(smile__tree_model_7I$finalModel)
smile__tree_model_7I_pred <- predict(smile__tree_model_7I, tst_smile)
summary(smile__tree_model_7I_pred)
## spontaneous deliberate
## 89 53
smile__tree_model_7I_confM <- confusionMatrix(
smile__tree_model_7I_pred,
tst_smile$smile_type
)
smile__tree_model_7I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 49 40
## deliberate 21 32
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.07664
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.02119
##
## Sensitivity : 0.7000
## Specificity : 0.4444
## Pos Pred Value : 0.5506
## Neg Pred Value : 0.6038
## Prevalence : 0.4930
## Detection Rate : 0.3451
## Detection Prevalence : 0.6268
## Balanced Accuracy : 0.5722
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7I.1_pred <- predict(smile__tree_model_7I, tst_smile_boys)
summary(smile__tree_model_7I.1_pred)
## spontaneous deliberate
## 41 36
smile__tree_model_7I.1_confM <- confusionMatrix(
smile__tree_model_7I.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 18
## deliberate 14 22
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1709
##
## Mcnemar's Test P-Value : 0.5959
##
## Sensitivity : 0.6216
## Specificity : 0.5500
## Pos Pred Value : 0.5610
## Neg Pred Value : 0.6111
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5325
## Balanced Accuracy : 0.5858
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7I.2_pred <- predict(smile__tree_model_7I, tst_smile_girls)
summary(smile__tree_model_7I.2_pred)
## spontaneous deliberate
## 48 17
smile__tree_model_7I.2_confM <- confusionMatrix(
smile__tree_model_7I.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 22
## deliberate 7 10
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1011
##
## Mcnemar's Test P-Value : 0.00933
##
## Sensitivity : 0.7879
## Specificity : 0.3125
## Pos Pred Value : 0.5417
## Neg Pred Value : 0.5882
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.7385
## Balanced Accuracy : 0.5502
##
## 'Positive' Class : spontaneous
##
# 7J apex
set.seed(1973)
smile__tree_model_7J <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
apex_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7J$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5454323 0.09550827 0.07304606 0.1452256
## 2 0.04545455 0.5393717 0.07923498 0.07276411 0.1460974
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7J$coefnames
## [1] "AU01_r_mean" "AU09_r_mean" "AU10_r_mean" "AU25_r_mean" "AU45_r_mean"
## [6] "apex_mean"
varImp(smile__tree_model_7J)
## rpart variable importance
##
## Overall
## AU25_r_mean 100.00
## AU45_r_mean 84.27
## AU01_r_mean 75.97
## AU09_r_mean 54.61
## apex_mean 27.02
## AU10_r_mean 0.00
# summary(smile__tree_model_7J$finalModel)
fancyRpartPlot(smile__tree_model_7J$finalModel)
smile__tree_model_7J_pred <- predict(smile__tree_model_7J, tst_smile)
summary(smile__tree_model_7J_pred)
## spontaneous deliberate
## 88 54
smile__tree_model_7J_confM <- confusionMatrix(
smile__tree_model_7J_pred,
tst_smile$smile_type
)
smile__tree_model_7J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 49 39
## deliberate 21 33
##
## Accuracy : 0.5775
## 95% CI : (0.4918, 0.6598)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.05517
##
## Kappa : 0.1578
##
## Mcnemar's Test P-Value : 0.02819
##
## Sensitivity : 0.7000
## Specificity : 0.4583
## Pos Pred Value : 0.5568
## Neg Pred Value : 0.6111
## Prevalence : 0.4930
## Detection Rate : 0.3451
## Detection Prevalence : 0.6197
## Balanced Accuracy : 0.5792
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7J.1_pred <- predict(smile__tree_model_7J, tst_smile_boys)
summary(smile__tree_model_7J.1_pred)
## spontaneous deliberate
## 39 38
smile__tree_model_7J.1_confM <- confusionMatrix(
smile__tree_model_7J.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 17
## deliberate 15 23
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1693
##
## Mcnemar's Test P-Value : 0.8597
##
## Sensitivity : 0.5946
## Specificity : 0.5750
## Pos Pred Value : 0.5641
## Neg Pred Value : 0.6053
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.5065
## Balanced Accuracy : 0.5848
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7J.2_pred <- predict(smile__tree_model_7J, tst_smile_girls)
summary(smile__tree_model_7J.2_pred)
## spontaneous deliberate
## 49 16
smile__tree_model_7J.2_confM <- confusionMatrix(
smile__tree_model_7J.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 27 22
## deliberate 6 10
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.192728
##
## Kappa : 0.1317
##
## Mcnemar's Test P-Value : 0.004586
##
## Sensitivity : 0.8182
## Specificity : 0.3125
## Pos Pred Value : 0.5510
## Neg Pred Value : 0.6250
## Prevalence : 0.5077
## Detection Rate : 0.4154
## Detection Prevalence : 0.7538
## Balanced Accuracy : 0.5653
##
## 'Positive' Class : spontaneous
##
# 7K
set.seed(1973)
smile__tree_model_7K <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
offset_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_7K$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03030303 0.5548017 0.11382011 0.08030484 0.1597152
## 2 0.04545455 0.5334893 0.06747028 0.07339185 0.1473916
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_7K$coefnames
## [1] "AU01_r_mean" "AU09_r_mean" "AU10_r_mean" "AU25_r_mean" "AU45_r_mean"
## [6] "offset_mean"
varImp(smile__tree_model_7K)
## rpart variable importance
##
## Overall
## AU25_r_mean 100.00
## AU45_r_mean 90.39
## AU01_r_mean 85.32
## AU10_r_mean 76.48
## AU09_r_mean 72.27
## offset_mean 0.00
# summary(smile__tree_model_7K$finalModel)
fancyRpartPlot(smile__tree_model_7K$finalModel)
smile__tree_model_7K_pred <- predict(smile__tree_model_7K, tst_smile)
summary(smile__tree_model_7K_pred)
## spontaneous deliberate
## 89 53
smile__tree_model_7K_confM <- confusionMatrix(
smile__tree_model_7K_pred,
tst_smile$smile_type
)
smile__tree_model_7K_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 49 40
## deliberate 21 32
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.07664
##
## Kappa : 0.1439
##
## Mcnemar's Test P-Value : 0.02119
##
## Sensitivity : 0.7000
## Specificity : 0.4444
## Pos Pred Value : 0.5506
## Neg Pred Value : 0.6038
## Prevalence : 0.4930
## Detection Rate : 0.3451
## Detection Prevalence : 0.6268
## Balanced Accuracy : 0.5722
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_7K.1_pred <- predict(smile__tree_model_7K, tst_smile_boys)
summary(smile__tree_model_7K.1_pred)
## spontaneous deliberate
## 41 36
smile__tree_model_7K.1_confM <- confusionMatrix(
smile__tree_model_7K.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_7K.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 18
## deliberate 14 22
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1709
##
## Mcnemar's Test P-Value : 0.5959
##
## Sensitivity : 0.6216
## Specificity : 0.5500
## Pos Pred Value : 0.5610
## Neg Pred Value : 0.6111
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5325
## Balanced Accuracy : 0.5858
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_7K.2_pred <- predict(smile__tree_model_7K, tst_smile_girls)
summary(smile__tree_model_7K.2_pred)
## spontaneous deliberate
## 48 17
smile__tree_model_7K.2_confM <- confusionMatrix(
smile__tree_model_7K.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_7K.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 22
## deliberate 7 10
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.26784
##
## Kappa : 0.1011
##
## Mcnemar's Test P-Value : 0.00933
##
## Sensitivity : 0.7879
## Specificity : 0.3125
## Pos Pred Value : 0.5417
## Neg Pred Value : 0.5882
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.7385
## Balanced Accuracy : 0.5502
##
## 'Positive' Class : spontaneous
##
# 8 strongest feature combination
# 8A
set.seed(1973)
smile__tree_model_8A <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean + lip_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8A$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5645443 0.13035315 0.04265332 0.08444669
## 2 0.05454545 0.5464516 0.09393859 0.04163563 0.08371676
## 3 0.18787879 0.5076649 0.01960290 0.05893775 0.12183508
smile__tree_model_8A$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "onset_mean" "apex_mean"
## [6] "offset_mean" "lip_mean" "eye_mean"
varImp(smile__tree_model_8A)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## offset_mean 69.75
## eye_mean 69.12
## AU06_r_mean 48.99
## onset_mean 26.75
## apex_mean 21.60
## AU12_r_mean 0.00
## lip_mean 0.00
# summary(smile__tree_model_8A$finalModel)
fancyRpartPlot(smile__tree_model_8A$finalModel)
smile__tree_model_8A_pred <- predict(smile__tree_model_8A, tst_smile)
summary(smile__tree_model_8A_pred)
## spontaneous deliberate
## 92 50
smile__tree_model_8A_confM <- confusionMatrix(
smile__tree_model_8A_pred,
tst_smile$smile_type
)
smile__tree_model_8A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 46
## deliberate 24 26
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.53358
##
## Kappa : 0.0182
##
## Mcnemar's Test P-Value : 0.01207
##
## Sensitivity : 0.6571
## Specificity : 0.3611
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5200
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.6479
## Balanced Accuracy : 0.5091
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8A.1_pred <- predict(smile__tree_model_8A, tst_smile_boys)
summary(smile__tree_model_8A.1_pred)
## spontaneous deliberate
## 49 28
smile__tree_model_8A.1_confM <- confusionMatrix(
smile__tree_model_8A.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 25
## deliberate 13 15
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.63425
##
## Kappa : 0.0234
##
## Mcnemar's Test P-Value : 0.07435
##
## Sensitivity : 0.6486
## Specificity : 0.3750
## Pos Pred Value : 0.4898
## Neg Pred Value : 0.5357
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.6364
## Balanced Accuracy : 0.5118
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8A.2_pred <- predict(smile__tree_model_8A, tst_smile_girls)
summary(smile__tree_model_8A.2_pred)
## spontaneous deliberate
## 43 22
smile__tree_model_8A.2_confM <- confusionMatrix(
smile__tree_model_8A.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 21
## deliberate 11 11
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0105
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.6667
## Specificity : 0.3438
## Pos Pred Value : 0.5116
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6615
## Balanced Accuracy : 0.5052
##
## 'Positive' Class : spontaneous
##
# 8B
set.seed(1973)
smile__tree_model_8B <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + lip_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8B$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5707832 0.14212731 0.03047154 0.06075247
## 2 0.06060606 0.5435105 0.09129408 0.03933110 0.08035001
## 3 0.18787879 0.5076649 0.01960290 0.05893775 0.12183508
smile__tree_model_8B$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "onset_mean" "lip_mean"
## [6] "eye_mean"
varImp(smile__tree_model_8B)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## eye_mean 69.12
## AU06_r_mean 48.99
## onset_mean 46.87
## lip_mean 27.55
## AU12_r_mean 0.00
# summary(smile__tree_model_8B$finalModel)
fancyRpartPlot(smile__tree_model_8B$finalModel)
smile__tree_model_8B_pred <- predict(smile__tree_model_8B, tst_smile)
summary(smile__tree_model_8B_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_8B_confM <- confusionMatrix(
smile__tree_model_8B_pred,
tst_smile$smile_type
)
smile__tree_model_8B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8B.1_pred <- predict(smile__tree_model_8B, tst_smile_boys)
summary(smile__tree_model_8B.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_8B.1_confM <- confusionMatrix(
smile__tree_model_8B.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8B.2_pred <- predict(smile__tree_model_8B, tst_smile_girls)
summary(smile__tree_model_8B.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_8B.2_confM <- confusionMatrix(
smile__tree_model_8B.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# 8C
set.seed(1973)
smile__tree_model_8C <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + apex_mean + lip_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8C$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5646335 0.13053791 0.02257622 0.04670496
## 2 0.06060606 0.5435105 0.09129408 0.03933110 0.08035001
## 3 0.18787879 0.5076649 0.01960290 0.05893775 0.12183508
smile__tree_model_8C$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "apex_mean" "lip_mean"
## [6] "eye_mean"
varImp(smile__tree_model_8C)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## eye_mean 69.12
## AU06_r_mean 48.99
## apex_mean 39.60
## lip_mean 27.55
## AU12_r_mean 0.00
# summary(smile__tree_model_8C$finalModel)
fancyRpartPlot(smile__tree_model_8C$finalModel)
smile__tree_model_8C_pred <- predict(smile__tree_model_8C, tst_smile)
summary(smile__tree_model_8C_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_8C_confM <- confusionMatrix(
smile__tree_model_8C_pred,
tst_smile$smile_type
)
smile__tree_model_8C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8C.1_pred <- predict(smile__tree_model_8C, tst_smile_boys)
summary(smile__tree_model_8C.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_8C.1_confM <- confusionMatrix(
smile__tree_model_8C.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8C.2_pred <- predict(smile__tree_model_8C, tst_smile_girls)
summary(smile__tree_model_8C.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_8C.2_confM <- confusionMatrix(
smile__tree_model_8C.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# 8D
set.seed(1973)
smile__tree_model_8D <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + offset_mean + lip_mean +
eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8D$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04848485 0.5525123 0.1060044 0.04568814 0.09216566
## 2 0.05454545 0.5525123 0.1060044 0.04568814 0.09216566
## 3 0.18787879 0.5076649 0.0196029 0.05893775 0.12183508
smile__tree_model_8D$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "offset_mean" "lip_mean"
## [6] "eye_mean"
varImp(smile__tree_model_8D)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## eye_mean 66.67
## offset_mean 50.87
## AU06_r_mean 38.96
## lip_mean 19.35
## AU12_r_mean 0.00
# summary(smile__tree_model_8D$finalModel)
fancyRpartPlot(smile__tree_model_8D$finalModel)
smile__tree_model_8D_pred <- predict(smile__tree_model_8D, tst_smile)
summary(smile__tree_model_8D_pred)
## spontaneous deliberate
## 102 40
smile__tree_model_8D_confM <- confusionMatrix(
smile__tree_model_8D_pred,
tst_smile$smile_type
)
smile__tree_model_8D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 52 50
## deliberate 18 22
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.4008030
##
## Kappa : 0.0481
##
## Mcnemar's Test P-Value : 0.0001704
##
## Sensitivity : 0.7429
## Specificity : 0.3056
## Pos Pred Value : 0.5098
## Neg Pred Value : 0.5500
## Prevalence : 0.4930
## Detection Rate : 0.3662
## Detection Prevalence : 0.7183
## Balanced Accuracy : 0.5242
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8D.1_pred <- predict(smile__tree_model_8D, tst_smile_boys)
summary(smile__tree_model_8D.1_pred)
## spontaneous deliberate
## 54 23
smile__tree_model_8D.1_confM <- confusionMatrix(
smile__tree_model_8D.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 27 27
## deliberate 10 13
##
## Accuracy : 0.5195
## 95% CI : (0.4026, 0.6348)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.545933
##
## Kappa : 0.0538
##
## Mcnemar's Test P-Value : 0.008529
##
## Sensitivity : 0.7297
## Specificity : 0.3250
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5652
## Prevalence : 0.4805
## Detection Rate : 0.3506
## Detection Prevalence : 0.7013
## Balanced Accuracy : 0.5274
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8D.2_pred <- predict(smile__tree_model_8D, tst_smile_girls)
summary(smile__tree_model_8D.2_pred)
## spontaneous deliberate
## 48 17
smile__tree_model_8D.2_confM <- confusionMatrix(
smile__tree_model_8D.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 23
## deliberate 8 9
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.45095
##
## Kappa : 0.0391
##
## Mcnemar's Test P-Value : 0.01192
##
## Sensitivity : 0.7576
## Specificity : 0.2812
## Pos Pred Value : 0.5208
## Neg Pred Value : 0.5294
## Prevalence : 0.5077
## Detection Rate : 0.3846
## Detection Prevalence : 0.7385
## Balanced Accuracy : 0.5194
##
## 'Positive' Class : spontaneous
##
# 8E
set.seed(1973)
smile__tree_model_8E <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean + lip_mean +
eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8E$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5335784 0.06951700 0.06719129 0.1351431
## 2 0.04545455 0.5394608 0.08128171 0.06932694 0.1391696
## 3 0.22424242 0.5255849 0.04429576 0.04913641 0.0997257
smile__tree_model_8E$coefnames
## [1] "AU01_r_mean" "AU09_r_mean" "AU10_r_mean" "AU25_r_mean" "AU45_r_mean"
## [6] "onset_mean" "apex_mean" "offset_mean" "lip_mean" "eye_mean"
varImp(smile__tree_model_8E)
## rpart variable importance
##
## Overall
## AU10_r_mean 100.00
## AU25_r_mean 92.60
## AU45_r_mean 82.12
## AU09_r_mean 62.58
## AU01_r_mean 60.90
## eye_mean 0.00
## lip_mean 0.00
## onset_mean 0.00
## apex_mean 0.00
## offset_mean 0.00
# summary(smile__tree_model_8E$finalModel)
fancyRpartPlot(smile__tree_model_8E$finalModel)
smile__tree_model_8E_pred <- predict(smile__tree_model_8E, tst_smile)
summary(smile__tree_model_8E_pred)
## spontaneous deliberate
## 41 101
smile__tree_model_8E_confM <- confusionMatrix(
smile__tree_model_8E_pred,
tst_smile$smile_type
)
smile__tree_model_8E_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 19
## deliberate 48 53
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3375712
##
## Kappa : 0.0507
##
## Mcnemar's Test P-Value : 0.0006245
##
## Sensitivity : 0.3143
## Specificity : 0.7361
## Pos Pred Value : 0.5366
## Neg Pred Value : 0.5248
## Prevalence : 0.4930
## Detection Rate : 0.1549
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5252
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8E.1_pred <- predict(smile__tree_model_8E, tst_smile_boys)
summary(smile__tree_model_8E.1_pred)
## spontaneous deliberate
## 13 64
smile__tree_model_8E.1_confM <- confusionMatrix(
smile__tree_model_8E.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 9 4
## deliberate 28 36
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1468
##
## Mcnemar's Test P-Value : 4.785e-05
##
## Sensitivity : 0.2432
## Specificity : 0.9000
## Pos Pred Value : 0.6923
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.1169
## Detection Prevalence : 0.1688
## Balanced Accuracy : 0.5716
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8E.2_pred <- predict(smile__tree_model_8E, tst_smile_girls)
summary(smile__tree_model_8E.2_pred)
## spontaneous deliberate
## 28 37
smile__tree_model_8E.2_confM <- confusionMatrix(
smile__tree_model_8E.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 15
## deliberate 20 17
##
## Accuracy : 0.4615
## 95% CI : (0.337, 0.5897)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.8074
##
## Kappa : -0.0746
##
## Mcnemar's Test P-Value : 0.4990
##
## Sensitivity : 0.3939
## Specificity : 0.5312
## Pos Pred Value : 0.4643
## Neg Pred Value : 0.4595
## Prevalence : 0.5077
## Detection Rate : 0.2000
## Detection Prevalence : 0.4308
## Balanced Accuracy : 0.4626
##
## 'Positive' Class : spontaneous
##
# 8F
set.seed(1973)
smile__tree_model_8F <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8F$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5584837 0.11790741 0.03491922 0.06900019
## 2 0.05454545 0.5464516 0.09393859 0.04163563 0.08371676
## 3 0.18787879 0.5076649 0.01960290 0.05893775 0.12183508
smile__tree_model_8F$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "onset_mean" "apex_mean"
## [6] "offset_mean" "eye_mean"
varImp(smile__tree_model_8F)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## offset_mean 69.75
## eye_mean 69.12
## AU06_r_mean 48.99
## onset_mean 26.75
## apex_mean 21.60
## AU12_r_mean 0.00
# summary(smile__tree_model_8F$finalModel)
fancyRpartPlot(smile__tree_model_8F$finalModel)
smile__tree_model_8F_pred <- predict(smile__tree_model_8F, tst_smile)
summary(smile__tree_model_8F_pred)
## spontaneous deliberate
## 92 50
smile__tree_model_8F_confM <- confusionMatrix(
smile__tree_model_8F_pred,
tst_smile$smile_type
)
smile__tree_model_8F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 46
## deliberate 24 26
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.53358
##
## Kappa : 0.0182
##
## Mcnemar's Test P-Value : 0.01207
##
## Sensitivity : 0.6571
## Specificity : 0.3611
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5200
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.6479
## Balanced Accuracy : 0.5091
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8F.1_pred <- predict(smile__tree_model_8F, tst_smile_boys)
summary(smile__tree_model_8F.1_pred)
## spontaneous deliberate
## 49 28
smile__tree_model_8F.1_confM <- confusionMatrix(
smile__tree_model_8F.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 25
## deliberate 13 15
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.63425
##
## Kappa : 0.0234
##
## Mcnemar's Test P-Value : 0.07435
##
## Sensitivity : 0.6486
## Specificity : 0.3750
## Pos Pred Value : 0.4898
## Neg Pred Value : 0.5357
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.6364
## Balanced Accuracy : 0.5118
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8F.2_pred <- predict(smile__tree_model_8F, tst_smile_girls)
summary(smile__tree_model_8F.2_pred)
## spontaneous deliberate
## 43 22
smile__tree_model_8F.2_confM <- confusionMatrix(
smile__tree_model_8F.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 21
## deliberate 11 11
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0105
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.6667
## Specificity : 0.3438
## Pos Pred Value : 0.5116
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6615
## Balanced Accuracy : 0.5052
##
## 'Positive' Class : spontaneous
##
# 8G
set.seed(1973)
smile__tree_model_8G <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8G$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5617814 0.12318742 0.03094505 0.06127823
## 2 0.06060606 0.5435105 0.09129408 0.03933110 0.08035001
## 3 0.18787879 0.5076649 0.01960290 0.05893775 0.12183508
smile__tree_model_8G$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "onset_mean" "eye_mean"
varImp(smile__tree_model_8G)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## eye_mean 58.79
## AU06_r_mean 31.92
## onset_mean 29.09
## AU12_r_mean 0.00
# summary(smile__tree_model_8G$finalModel)
fancyRpartPlot(smile__tree_model_8G$finalModel)
smile__tree_model_8G_pred <- predict(smile__tree_model_8G, tst_smile)
summary(smile__tree_model_8G_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_8G_confM <- confusionMatrix(
smile__tree_model_8G_pred,
tst_smile$smile_type
)
smile__tree_model_8G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8G.1_pred <- predict(smile__tree_model_8G, tst_smile_boys)
summary(smile__tree_model_8G.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_8G.1_confM <- confusionMatrix(
smile__tree_model_8G.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8G.2_pred <- predict(smile__tree_model_8G, tst_smile_girls)
summary(smile__tree_model_8G.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_8G.2_confM <- confusionMatrix(
smile__tree_model_8G.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# 8H apex
set.seed(1973)
smile__tree_model_8H <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + apex_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8H$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04242424 0.5647226 0.12902381 0.03522295 0.07058369
## 2 0.06060606 0.5435105 0.09129408 0.03933110 0.08035001
## 3 0.18787879 0.5076649 0.01960290 0.05893775 0.12183508
smile__tree_model_8H$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "apex_mean" "eye_mean"
varImp(smile__tree_model_8H)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## eye_mean 58.79
## AU06_r_mean 31.92
## apex_mean 19.39
## AU12_r_mean 0.00
# summary(smile__tree_model_8H$finalModel)
fancyRpartPlot(smile__tree_model_8H$finalModel)
smile__tree_model_8H_pred <- predict(smile__tree_model_8H, tst_smile)
summary(smile__tree_model_8H_pred)
## spontaneous deliberate
## 85 57
smile__tree_model_8H_confM <- confusionMatrix(
smile__tree_model_8H_pred,
tst_smile$smile_type
)
smile__tree_model_8H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 41
## deliberate 26 31
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.059
##
## Mcnemar's Test P-Value : 0.0872
##
## Sensitivity : 0.6286
## Specificity : 0.4306
## Pos Pred Value : 0.5176
## Neg Pred Value : 0.5439
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5986
## Balanced Accuracy : 0.5296
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8H.1_pred <- predict(smile__tree_model_8H, tst_smile_boys)
summary(smile__tree_model_8H.1_pred)
## spontaneous deliberate
## 45 32
smile__tree_model_8H.1_confM <- confusionMatrix(
smile__tree_model_8H.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 22
## deliberate 14 18
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.071
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.6216
## Specificity : 0.4500
## Pos Pred Value : 0.5111
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5844
## Balanced Accuracy : 0.5358
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8H.2_pred <- predict(smile__tree_model_8H, tst_smile_girls)
summary(smile__tree_model_8H.2_pred)
## spontaneous deliberate
## 40 25
smile__tree_model_8H.2_confM <- confusionMatrix(
smile__tree_model_8H.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 19
## deliberate 12 13
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0428
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.6364
## Specificity : 0.4062
## Pos Pred Value : 0.5250
## Neg Pred Value : 0.5200
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5213
##
## 'Positive' Class : spontaneous
##
# 8I offset
set.seed(1973)
smile__tree_model_8I <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + offset_mean + eye_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8I$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.04848485 0.5494820 0.10103869 0.04264550 0.08677768
## 2 0.05454545 0.5464516 0.09556325 0.04163563 0.08389847
## 3 0.18787879 0.5076649 0.01960290 0.05893775 0.12183508
smile__tree_model_8I$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "offset_mean" "eye_mean"
varImp(smile__tree_model_8I)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## offset_mean 59.63
## eye_mean 58.79
## AU06_r_mean 31.92
## AU12_r_mean 0.00
# summary(smile__tree_model_8I$finalModel)
fancyRpartPlot(smile__tree_model_8I$finalModel)
smile__tree_model_8I_pred <- predict(smile__tree_model_8I, tst_smile)
summary(smile__tree_model_8I_pred)
## spontaneous deliberate
## 92 50
smile__tree_model_8I_confM <- confusionMatrix(
smile__tree_model_8I_pred,
tst_smile$smile_type
)
smile__tree_model_8I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 46
## deliberate 24 26
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.53358
##
## Kappa : 0.0182
##
## Mcnemar's Test P-Value : 0.01207
##
## Sensitivity : 0.6571
## Specificity : 0.3611
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5200
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.6479
## Balanced Accuracy : 0.5091
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8I.1_pred <- predict(smile__tree_model_8I, tst_smile_boys)
summary(smile__tree_model_8I.1_pred)
## spontaneous deliberate
## 49 28
smile__tree_model_8I.1_confM <- confusionMatrix(
smile__tree_model_8I.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 25
## deliberate 13 15
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.63425
##
## Kappa : 0.0234
##
## Mcnemar's Test P-Value : 0.07435
##
## Sensitivity : 0.6486
## Specificity : 0.3750
## Pos Pred Value : 0.4898
## Neg Pred Value : 0.5357
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.6364
## Balanced Accuracy : 0.5118
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8I.2_pred <- predict(smile__tree_model_8I, tst_smile_girls)
summary(smile__tree_model_8I.2_pred)
## spontaneous deliberate
## 43 22
smile__tree_model_8I.2_confM <- confusionMatrix(
smile__tree_model_8I.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 21
## deliberate 11 11
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0105
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.6667
## Specificity : 0.3438
## Pos Pred Value : 0.5116
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6615
## Balanced Accuracy : 0.5052
##
## 'Positive' Class : spontaneous
##
# 8J
set.seed(1973)
smile__tree_model_8J <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean + lip_mean,
method = "rpart", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__tree_model_8J$results
## cp Accuracy Kappa AccuracySD KappaSD
## 1 0.03636364 0.5704267 0.14211785 0.04974950 0.09859519
## 2 0.05454545 0.5493928 0.09982095 0.04575070 0.09183336
## 3 0.18787879 0.5223708 0.04901466 0.03672728 0.07788091
smile__tree_model_8J$coefnames
## [1] "AU06_r_mean" "AU12_r_mean" "AU45_r_mean" "onset_mean" "apex_mean"
## [6] "offset_mean" "lip_mean"
varImp(smile__tree_model_8J)
## rpart variable importance
##
## Overall
## AU45_r_mean 100.00
## offset_mean 69.75
## AU06_r_mean 48.99
## onset_mean 46.87
## apex_mean 39.60
## AU12_r_mean 0.00
## lip_mean 0.00
# summary(smile__tree_model_8J$finalModel)
fancyRpartPlot(smile__tree_model_8J$finalModel)
smile__tree_model_8J_pred <- predict(smile__tree_model_8J, tst_smile)
summary(smile__tree_model_8J_pred)
## spontaneous deliberate
## 92 50
smile__tree_model_8J_confM <- confusionMatrix(
smile__tree_model_8J_pred,
tst_smile$smile_type
)
smile__tree_model_8J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 46
## deliberate 24 26
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.53358
##
## Kappa : 0.0182
##
## Mcnemar's Test P-Value : 0.01207
##
## Sensitivity : 0.6571
## Specificity : 0.3611
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5200
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.6479
## Balanced Accuracy : 0.5091
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__tree_model_8J.1_pred <- predict(smile__tree_model_8J, tst_smile_boys)
summary(smile__tree_model_8J.1_pred)
## spontaneous deliberate
## 49 28
smile__tree_model_8J.1_confM <- confusionMatrix(
smile__tree_model_8J.1_pred,
tst_smile_boys$smile_type
)
smile__tree_model_8J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 25
## deliberate 13 15
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.63425
##
## Kappa : 0.0234
##
## Mcnemar's Test P-Value : 0.07435
##
## Sensitivity : 0.6486
## Specificity : 0.3750
## Pos Pred Value : 0.4898
## Neg Pred Value : 0.5357
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.6364
## Balanced Accuracy : 0.5118
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__tree_model_8J.2_pred <- predict(smile__tree_model_8J, tst_smile_girls)
summary(smile__tree_model_8J.2_pred)
## spontaneous deliberate
## 43 22
smile__tree_model_8J.2_confM <- confusionMatrix(
smile__tree_model_8J.2_pred,
tst_smile_girls$smile_type
)
smile__tree_model_8J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 21
## deliberate 11 11
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0105
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.6667
## Specificity : 0.3438
## Pos Pred Value : 0.5116
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6615
## Balanced Accuracy : 0.5052
##
## 'Positive' Class : spontaneous
##
For SVM the kernlab package is used. The method used in the caret package is svmlinear. More information about the ksvm packages can be found in the citation link or ? R help function. The trained models are divided into the same eight categories as decision trees. Two additional strong models were build for SVM. Again, multiple models per category are explored. The explanation on the categories can be found in the thesis. To train the models the train() function is used. The models are stored as variable. The parameter settings are explained in the thesis. The models use 10 fold cross-validation. The same pre-processing check is added to the complete model. On this first complete model, a ROC example has been visualized. As this is not an evaluation parameter in the thesis, it is not used any further. The cost function (C) is set to \(1\) which is the default in for this model. To visualize the trained SVM based on two features the kernlab plot is used. Some examples are displayed in the code. The predict() function is used to create the predictions based on the test set, and stored as variable. For model evaluation the confusionMatrix() function is used and printed.
# load packages
library(kernlab)
# set seed
set.seed(1973)
# complete model 0 with cost function set to 1 (default)
smile__svm_model_0 <- train(smile_type ~ .,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
# check different parameters of the model results
smile__svm_model_0
## Support Vector Machines with Linear Kernel
##
## 333 samples
## 32 predictor
## 2 classes: 'spontaneous', 'deliberate '
##
## No pre-processing
## Resampling: Cross-Validated (10 fold)
## Summary of sample sizes: 300, 300, 300, 299, 299, 300, ...
## Resampling results:
##
## Accuracy Kappa
## 0.7050914 0.4098127
##
## Tuning parameter 'C' was held constant at a value of 1
smile__svm_model_0$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.7050914 0.4098127 0.1064684 0.2126184
smile__svm_model_0$resample
## Accuracy Kappa Resample
## 1 0.8181818 0.63602941 Fold01
## 2 0.7878788 0.57458564 Fold02
## 3 0.6666667 0.33394495 Fold03
## 4 0.7941176 0.58823529 Fold04
## 5 0.5294118 0.05882353 Fold05
## 6 0.7272727 0.45101664 Fold06
## 7 0.6060606 0.21284404 Fold07
## 8 0.7941176 0.58823529 Fold08
## 9 0.5625000 0.12500000 Fold09
## 10 0.7647059 0.52941176 Fold10
smile__svm_model_0$bestTune
## C
## 1 1
# summary(smile__svm_model_0$finalModel) - not printed
# prediction on the test set and the model
smile__svm_model_0_pred <- predict(smile__svm_model_0, newdata = tst_smile)
# print prediction
summary(smile__svm_model_0_pred)
## spontaneous deliberate
## 67 75
# Evaluation of the accuracy based on the confusion matrix
smile__svm_model_0_confM <- confusionMatrix(
smile__svm_model_0_pred,
tst_smile$smile_type
)
# print the confusion matrix
smile__svm_model_0_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 50 17
## deliberate 20 55
##
## Accuracy : 0.7394
## 95% CI : (0.6592, 0.8094)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 1.288e-08
##
## Kappa : 0.4785
##
## Mcnemar's Test P-Value : 0.7423
##
## Sensitivity : 0.7143
## Specificity : 0.7639
## Pos Pred Value : 0.7463
## Neg Pred Value : 0.7333
## Prevalence : 0.4930
## Detection Rate : 0.3521
## Detection Prevalence : 0.4718
## Balanced Accuracy : 0.7391
##
## 'Positive' Class : spontaneous
##
# set seed
set.seed(1973)
# complete model 0 + pre-processing - outcome does not improve!
smile__svm_model_0.0.1 <- train(smile_type ~ .,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10),
preProcess = c("center", "scale")
)
smile__svm_model_0.0.1
## Support Vector Machines with Linear Kernel
##
## 333 samples
## 32 predictor
## 2 classes: 'spontaneous', 'deliberate '
##
## Pre-processing: centered (32), scaled (32)
## Resampling: Cross-Validated (10 fold)
## Summary of sample sizes: 300, 300, 300, 299, 299, 300, ...
## Resampling results:
##
## Accuracy Kappa
## 0.7050914 0.4098127
##
## Tuning parameter 'C' was held constant at a value of 1
smile__svm_model_0.0.1$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.7050914 0.4098127 0.1064684 0.2126184
# predicting boys, girls
set.seed(1973)
smile__svm_model_0.1_pred <- predict(smile__svm_model_0, tst_smile_boys)
summary(smile__svm_model_0.1_pred)
## spontaneous deliberate
## 33 44
smile__svm_model_0.1_confM <- confusionMatrix(
smile__svm_model_0.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_0.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 7
## deliberate 11 33
##
## Accuracy : 0.7662
## 95% CI : (0.6559, 0.8552)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 7.359e-06
##
## Kappa : 0.5299
##
## Mcnemar's Test P-Value : 0.4795
##
## Sensitivity : 0.7027
## Specificity : 0.8250
## Pos Pred Value : 0.7879
## Neg Pred Value : 0.7500
## Prevalence : 0.4805
## Detection Rate : 0.3377
## Detection Prevalence : 0.4286
## Balanced Accuracy : 0.7639
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_0.2_pred <- predict(smile__svm_model_0, tst_smile_girls)
summary(smile__svm_model_0.2_pred)
## spontaneous deliberate
## 34 31
smile__svm_model_0.2_confM <- confusionMatrix(
smile__svm_model_0.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_0.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 10
## deliberate 9 22
##
## Accuracy : 0.7077
## 95% CI : (0.5817, 0.814)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.0008348
##
## Kappa : 0.415
##
## Mcnemar's Test P-Value : 1.0000000
##
## Sensitivity : 0.7273
## Specificity : 0.6875
## Pos Pred Value : 0.7059
## Neg Pred Value : 0.7097
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.5231
## Balanced Accuracy : 0.7074
##
## 'Positive' Class : spontaneous
##
# example ROC
# load packages
library(pROC)
# citation("pROC")
# ROC model 0
roc_0 <- roc(
as.numeric(tst_smile$smile_type),
as.numeric(as.factor(smile__svm_model_0_pred))
)
roc_0
##
## Call:
## roc.default(response = as.numeric(tst_smile$smile_type), predictor = as.numeric(as.factor(smile__svm_model_0_pred)))
##
## Data: as.numeric(as.factor(smile__svm_model_0_pred)) in 70 controls (as.numeric(tst_smile$smile_type) 1) < 72 cases (as.numeric(tst_smile$smile_type) 2).
## Area under the curve: 0.7391
# Visualize the ROC model
par(mfrow = c(1, 1))
par(mai = c(.9, .8, .2, .2))
plot.roc(roc_0,
print.auc = TRUE, col = "black", lwd = 1,
main = "ROC curve", xlab = "Specificity: true negative rate",
ylab = "Sensitivity: true positive rate",
xlim = c(1, 0), ylim = c(0, 1), print.thres = "best"
)
abline(v = 1, lty = 2)
abline(h = 1, lty = 2)
text(.90, .97, labels = "Ideal Model")
points(1, 1, pch = "O", cex = 1.5)
# citation("kernlab")
# model 1 onset-apex-offset
set.seed(1973)
smile__svm_model_1 <- train(smile_type ~ onset_mean + offset_mean + apex_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_1$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6448362 0.2904029 0.1026657 0.2039863
smile__svm_model_1_pred <- predict(smile__svm_model_1, newdata = tst_smile)
summary(smile__svm_model_1_pred)
## spontaneous deliberate
## 76 66
smile__svm_model_1_confM <- confusionMatrix(
smile__svm_model_1_pred,
tst_smile$smile_type
)
smile__svm_model_1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 47 29
## deliberate 23 43
##
## Accuracy : 0.6338
## 95% CI : (0.5489, 0.713)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.001568
##
## Kappa : 0.2683
##
## Mcnemar's Test P-Value : 0.488074
##
## Sensitivity : 0.6714
## Specificity : 0.5972
## Pos Pred Value : 0.6184
## Neg Pred Value : 0.6515
## Prevalence : 0.4930
## Detection Rate : 0.3310
## Detection Prevalence : 0.5352
## Balanced Accuracy : 0.6343
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_1.1_pred <- predict(smile__svm_model_1, tst_smile_boys)
summary(smile__svm_model_1.1_pred)
## spontaneous deliberate
## 38 39
smile__svm_model_1.1_confM <- confusionMatrix(
smile__svm_model_1.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_1.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 15
## deliberate 14 25
##
## Accuracy : 0.6234
## 95% CI : (0.5056, 0.7313)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.04297
##
## Kappa : 0.2464
##
## Mcnemar's Test P-Value : 1.00000
##
## Sensitivity : 0.6216
## Specificity : 0.6250
## Pos Pred Value : 0.6053
## Neg Pred Value : 0.6410
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.4935
## Balanced Accuracy : 0.6233
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_1.2_pred <- predict(smile__svm_model_1, tst_smile_girls)
summary(smile__svm_model_1.2_pred)
## spontaneous deliberate
## 38 27
smile__svm_model_1.2_confM <- confusionMatrix(
smile__svm_model_1.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_1.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 14
## deliberate 9 18
##
## Accuracy : 0.6462
## 95% CI : (0.5177, 0.7608)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.01698
##
## Kappa : 0.2905
##
## Mcnemar's Test P-Value : 0.40425
##
## Sensitivity : 0.7273
## Specificity : 0.5625
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.5846
## Balanced Accuracy : 0.6449
##
## 'Positive' Class : spontaneous
##
# Roc model 1
roc_1 <- roc(
as.numeric(tst_smile$smile_type),
as.numeric(as.factor(smile__svm_model_1_pred))
)
roc_1
##
## Call:
## roc.default(response = as.numeric(tst_smile$smile_type), predictor = as.numeric(as.factor(smile__svm_model_1_pred)))
##
## Data: as.numeric(as.factor(smile__svm_model_1_pred)) in 70 controls (as.numeric(tst_smile$smile_type) 1) < 72 cases (as.numeric(tst_smile$smile_type) 2).
## Area under the curve: 0.6343
par(mfrow = c(1, 1))
par(mai = c(.9, .8, .2, .2))
plot.roc(roc_1,
print.auc = TRUE, col = "black", lwd = 1,
main = "ROC curve", xlab = "Specificity: true negative rate",
ylab = "Sensitivity: true positive rate",
xlim = c(1, 0), ylim = c(0, 1), print.thres = "best"
)
abline(v = 1, lty = 2)
abline(h = 1, lty = 2)
text(.90, .97, labels = "Ideal Model")
points(1, 1, pch = "O", cex = 1.5)
# model 1A onset
set.seed(1973)
smile__svm_model_1A <- train(smile_type ~ onset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_1A$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.4977607 -0.0005964829 0.07121788 0.1409985
smile__svm_model_1A_pred <- predict(smile__svm_model_1A, newdata = tst_smile)
summary(smile__svm_model_1A_pred)
## spontaneous deliberate
## 72 70
smile__svm_model_1A_confM <- confusionMatrix(
smile__svm_model_1A_pred,
tst_smile$smile_type
)
smile__svm_model_1A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 42 30
## deliberate 28 42
##
## Accuracy : 0.5915
## 95% CI : (0.506, 0.6732)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.02653
##
## Kappa : 0.1833
##
## Mcnemar's Test P-Value : 0.89553
##
## Sensitivity : 0.6000
## Specificity : 0.5833
## Pos Pred Value : 0.5833
## Neg Pred Value : 0.6000
## Prevalence : 0.4930
## Detection Rate : 0.2958
## Detection Prevalence : 0.5070
## Balanced Accuracy : 0.5917
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_1A.1_pred <- predict(smile__svm_model_1A, tst_smile_boys)
summary(smile__svm_model_1A.1_pred)
## spontaneous deliberate
## 37 40
smile__svm_model_1A.1_confM <- confusionMatrix(
smile__svm_model_1A.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_1A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 16
## deliberate 16 24
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1676
##
## Mcnemar's Test P-Value : 1.0000
##
## Sensitivity : 0.5676
## Specificity : 0.6000
## Pos Pred Value : 0.5676
## Neg Pred Value : 0.6000
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.4805
## Balanced Accuracy : 0.5838
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_1A.2_pred <- predict(smile__svm_model_1A, tst_smile_girls)
summary(smile__svm_model_1A.2_pred)
## spontaneous deliberate
## 35 30
smile__svm_model_1A.2_confM <- confusionMatrix(
smile__svm_model_1A.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_1A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 14
## deliberate 12 18
##
## Accuracy : 0.6
## 95% CI : (0.471, 0.7196)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.08588
##
## Kappa : 0.1991
##
## Mcnemar's Test P-Value : 0.84452
##
## Sensitivity : 0.6364
## Specificity : 0.5625
## Pos Pred Value : 0.6000
## Neg Pred Value : 0.6000
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.5385
## Balanced Accuracy : 0.5994
##
## 'Positive' Class : spontaneous
##
# model 1B apex
set.seed(1973)
smile__svm_model_1B <- train(smile_type ~ apex_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_1B$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5408422 0.0803652 0.07503868 0.1492659
smile__svm_model_1B_pred <- predict(smile__svm_model_1B, newdata = tst_smile)
summary(smile__svm_model_1B_pred)
## spontaneous deliberate
## 57 85
smile__svm_model_1B_confM <- confusionMatrix(
smile__svm_model_1B_pred,
tst_smile$smile_type
)
smile__svm_model_1B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 31
## deliberate 44 41
##
## Accuracy : 0.4718
## 95% CI : (0.3876, 0.5573)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.8220
##
## Kappa : -0.0593
##
## Mcnemar's Test P-Value : 0.1659
##
## Sensitivity : 0.3714
## Specificity : 0.5694
## Pos Pred Value : 0.4561
## Neg Pred Value : 0.4824
## Prevalence : 0.4930
## Detection Rate : 0.1831
## Detection Prevalence : 0.4014
## Balanced Accuracy : 0.4704
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_1B.1_pred <- predict(smile__svm_model_1B, tst_smile_boys)
summary(smile__svm_model_1B.1_pred)
## spontaneous deliberate
## 31 46
smile__svm_model_1B.1_confM <- confusionMatrix(
smile__svm_model_1B.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_1B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 15 16
## deliberate 22 24
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.6342
##
## Kappa : 0.0054
##
## Mcnemar's Test P-Value : 0.4173
##
## Sensitivity : 0.4054
## Specificity : 0.6000
## Pos Pred Value : 0.4839
## Neg Pred Value : 0.5217
## Prevalence : 0.4805
## Detection Rate : 0.1948
## Detection Prevalence : 0.4026
## Balanced Accuracy : 0.5027
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_1B.2_pred <- predict(smile__svm_model_1B, tst_smile_girls)
summary(smile__svm_model_1B.2_pred)
## spontaneous deliberate
## 26 39
smile__svm_model_1B.2_confM <- confusionMatrix(
smile__svm_model_1B.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_1B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 11 15
## deliberate 22 17
##
## Accuracy : 0.4308
## 95% CI : (0.3085, 0.5596)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.9139
##
## Kappa : -0.135
##
## Mcnemar's Test P-Value : 0.3239
##
## Sensitivity : 0.3333
## Specificity : 0.5312
## Pos Pred Value : 0.4231
## Neg Pred Value : 0.4359
## Prevalence : 0.5077
## Detection Rate : 0.1692
## Detection Prevalence : 0.4000
## Balanced Accuracy : 0.4323
##
## 'Positive' Class : spontaneous
##
# model 1C offset
set.seed(1973)
smile__svm_model_1C <- train(smile_type ~ offset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_1C$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5464294 0.09219777 0.05082367 0.1022998
smile__svm_model_1C_pred <- predict(smile__svm_model_1C, newdata = tst_smile)
summary(smile__svm_model_1C_pred)
## spontaneous deliberate
## 76 66
smile__svm_model_1C_confM <- confusionMatrix(
smile__svm_model_1C_pred,
tst_smile$smile_type
)
smile__svm_model_1C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 45 31
## deliberate 25 41
##
## Accuracy : 0.6056
## 95% CI : (0.5202, 0.6865)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.01151
##
## Kappa : 0.212
##
## Mcnemar's Test P-Value : 0.50404
##
## Sensitivity : 0.6429
## Specificity : 0.5694
## Pos Pred Value : 0.5921
## Neg Pred Value : 0.6212
## Prevalence : 0.4930
## Detection Rate : 0.3169
## Detection Prevalence : 0.5352
## Balanced Accuracy : 0.6062
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_1C.1_pred <- predict(smile__svm_model_1C, tst_smile_boys)
summary(smile__svm_model_1C.1_pred)
## spontaneous deliberate
## 40 37
smile__svm_model_1C.1_confM <- confusionMatrix(
smile__svm_model_1C.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_1C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 18
## deliberate 15 22
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.2126
##
## Kappa : 0.1442
##
## Mcnemar's Test P-Value : 0.7277
##
## Sensitivity : 0.5946
## Specificity : 0.5500
## Pos Pred Value : 0.5500
## Neg Pred Value : 0.5946
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.5195
## Balanced Accuracy : 0.5723
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_1C.2_pred <- predict(smile__svm_model_1C, tst_smile_girls)
summary(smile__svm_model_1C.2_pred)
## spontaneous deliberate
## 36 29
smile__svm_model_1C.2_confM <- confusionMatrix(
smile__svm_model_1C.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_1C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 13
## deliberate 10 19
##
## Accuracy : 0.6462
## 95% CI : (0.5177, 0.7608)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.01698
##
## Kappa : 0.2911
##
## Mcnemar's Test P-Value : 0.67666
##
## Sensitivity : 0.6970
## Specificity : 0.5938
## Pos Pred Value : 0.6389
## Neg Pred Value : 0.6552
## Prevalence : 0.5077
## Detection Rate : 0.3538
## Detection Prevalence : 0.5538
## Balanced Accuracy : 0.6454
##
## 'Positive' Class : spontaneous
##
# model 2 complete excluding subject info
set.seed(1973)
smile__svm_model_2 <- train(smile_type ~ . - subject - age,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10),
)
smile__svm_model_2$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.728799 0.4571015 0.1104272 0.2212161
smile__svm_model_2_pred <- predict(smile__svm_model_2, tst_smile)
summary(smile__svm_model_2_pred)
## spontaneous deliberate
## 61 81
smile__tree_svm_2_confM <- confusionMatrix(
smile__svm_model_2_pred,
tst_smile$smile_type
)
smile__tree_svm_2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 17
## deliberate 26 55
##
## Accuracy : 0.6972
## 95% CI : (0.6145, 0.7714)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 3.274e-06
##
## Kappa : 0.3932
##
## Mcnemar's Test P-Value : 0.2225
##
## Sensitivity : 0.6286
## Specificity : 0.7639
## Pos Pred Value : 0.7213
## Neg Pred Value : 0.6790
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.4296
## Balanced Accuracy : 0.6962
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_2.1_pred <- predict(smile__svm_model_2, tst_smile_boys)
summary(smile__svm_model_2.1_pred)
## spontaneous deliberate
## 28 49
smile__svm_model_2.1_confM <- confusionMatrix(
smile__svm_model_2.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_2.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 6
## deliberate 15 34
##
## Accuracy : 0.7273
## 95% CI : (0.6138, 0.8226)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.0001583
##
## Kappa : 0.4487
##
## Mcnemar's Test P-Value : 0.0808556
##
## Sensitivity : 0.5946
## Specificity : 0.8500
## Pos Pred Value : 0.7857
## Neg Pred Value : 0.6939
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.3636
## Balanced Accuracy : 0.7223
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_2.2_pred <- predict(smile__svm_model_2, tst_smile_girls)
summary(smile__svm_model_2.2_pred)
## spontaneous deliberate
## 33 32
smile__svm_model_2.2_confM <- confusionMatrix(
smile__svm_model_2.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_2.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 11
## deliberate 11 21
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3229
##
## Mcnemar's Test P-Value : 1.000000
##
## Sensitivity : 0.6667
## Specificity : 0.6562
## Pos Pred Value : 0.6667
## Neg Pred Value : 0.6562
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.5077
## Balanced Accuracy : 0.6615
##
## 'Positive' Class : spontaneous
##
# model 3 complete lip and eye features
set.seed(1973)
smile__svm_model_3 <- train(smile_type ~ lip_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_3$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5143661 0.02580907 0.08044698 0.1609386
summary(smile__svm_model_3$finalModel)
## Length Class Mode
## 1 ksvm S4
kernlab::plot(smile__svm_model_3$finalModel)
smile__svm_model_3_pred <- predict(smile__svm_model_3, tst_smile)
summary(smile__svm_model_3_pred)
## spontaneous deliberate
## 55 87
smile__svm_model_3_confM <- confusionMatrix(
smile__svm_model_3_pred,
tst_smile$smile_type
)
smile__svm_model_3_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 31 24
## deliberate 39 48
##
## Accuracy : 0.5563
## 95% CI : (0.4707, 0.6396)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.13758
##
## Kappa : 0.1099
##
## Mcnemar's Test P-Value : 0.07776
##
## Sensitivity : 0.4429
## Specificity : 0.6667
## Pos Pred Value : 0.5636
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.2183
## Detection Prevalence : 0.3873
## Balanced Accuracy : 0.5548
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_3.1_pred <- predict(smile__svm_model_3, tst_smile_boys)
summary(smile__svm_model_3.1_pred)
## spontaneous deliberate
## 28 49
smile__svm_model_3.1_confM <- confusionMatrix(
smile__svm_model_3.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_3.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 17 11
## deliberate 20 29
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.1862
##
## Mcnemar's Test P-Value : 0.1508
##
## Sensitivity : 0.4595
## Specificity : 0.7250
## Pos Pred Value : 0.6071
## Neg Pred Value : 0.5918
## Prevalence : 0.4805
## Detection Rate : 0.2208
## Detection Prevalence : 0.3636
## Balanced Accuracy : 0.5922
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_3.2_pred <- predict(smile__svm_model_3, tst_smile_girls)
summary(smile__svm_model_3.2_pred)
## spontaneous deliberate
## 27 38
smile__svm_model_3.2_confM <- confusionMatrix(
smile__svm_model_3.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_3.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 14 13
## deliberate 19 19
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0179
##
## Mcnemar's Test P-Value : 0.3768
##
## Sensitivity : 0.4242
## Specificity : 0.5938
## Pos Pred Value : 0.5185
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.2154
## Detection Prevalence : 0.4154
## Balanced Accuracy : 0.5090
##
## 'Positive' Class : spontaneous
##
# model 3A complete lip and eye features
set.seed(1973)
smile__svm_model_3A <- train(smile_type ~ lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_3A$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5262255 0.05020194 0.08590419 0.1693449
summary(smile__svm_model_3A$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_3A_pred <- predict(smile__svm_model_3A, tst_smile)
summary(smile__svm_model_3A_pred)
## spontaneous deliberate
## 46 96
smile__svm_model_3A_confM <- confusionMatrix(
smile__svm_model_3A_pred,
tst_smile$smile_type
)
smile__svm_model_3A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 23
## deliberate 47 49
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.533579
##
## Kappa : 0.0092
##
## Mcnemar's Test P-Value : 0.005977
##
## Sensitivity : 0.3286
## Specificity : 0.6806
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5104
## Prevalence : 0.4930
## Detection Rate : 0.1620
## Detection Prevalence : 0.3239
## Balanced Accuracy : 0.5046
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_3A.1_pred <- predict(smile__svm_model_3A, tst_smile_boys)
summary(smile__svm_model_3A.1_pred)
## spontaneous deliberate
## 24 53
smile__svm_model_3A.1_confM <- confusionMatrix(
smile__svm_model_3A.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_3A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 11
## deliberate 24 29
##
## Accuracy : 0.5455
## 95% CI : (0.4279, 0.6594)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.36674
##
## Kappa : 0.0774
##
## Mcnemar's Test P-Value : 0.04252
##
## Sensitivity : 0.3514
## Specificity : 0.7250
## Pos Pred Value : 0.5417
## Neg Pred Value : 0.5472
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.3117
## Balanced Accuracy : 0.5382
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_3A.2_pred <- predict(smile__svm_model_3A, tst_smile_girls)
summary(smile__svm_model_3A.2_pred)
## spontaneous deliberate
## 22 43
smile__svm_model_3A.2_confM <- confusionMatrix(
smile__svm_model_3A.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_3A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 10 12
## deliberate 23 20
##
## Accuracy : 0.4615
## 95% CI : (0.337, 0.5897)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.80736
##
## Kappa : -0.0716
##
## Mcnemar's Test P-Value : 0.09097
##
## Sensitivity : 0.3030
## Specificity : 0.6250
## Pos Pred Value : 0.4545
## Neg Pred Value : 0.4651
## Prevalence : 0.5077
## Detection Rate : 0.1538
## Detection Prevalence : 0.3385
## Balanced Accuracy : 0.4640
##
## 'Positive' Class : spontaneous
##
# model 3B eye
set.seed(1973)
smile__svm_model_3B <- train(smile_type ~ eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_3B$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5529523 0.1030544 0.09511338 0.1911904
summary(smile__svm_model_3B$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_3B_pred <- predict(smile__svm_model_3B, tst_smile)
summary(smile__svm_model_3B_pred)
## spontaneous deliberate
## 50 92
smile__svm_model_3B_confM <- confusionMatrix(
smile__svm_model_3B_pred,
tst_smile$smile_type
)
smile__svm_model_3B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 30 20
## deliberate 40 52
##
## Accuracy : 0.5775
## 95% CI : (0.4918, 0.6598)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.05517
##
## Kappa : 0.1514
##
## Mcnemar's Test P-Value : 0.01417
##
## Sensitivity : 0.4286
## Specificity : 0.7222
## Pos Pred Value : 0.6000
## Neg Pred Value : 0.5652
## Prevalence : 0.4930
## Detection Rate : 0.2113
## Detection Prevalence : 0.3521
## Balanced Accuracy : 0.5754
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_3B.1_pred <- predict(smile__svm_model_3B, tst_smile_boys)
summary(smile__svm_model_3B.1_pred)
## spontaneous deliberate
## 26 51
smile__svm_model_3B.1_confM <- confusionMatrix(
smile__svm_model_3B.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_3B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 10
## deliberate 21 30
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.10453
##
## Kappa : 0.1845
##
## Mcnemar's Test P-Value : 0.07249
##
## Sensitivity : 0.4324
## Specificity : 0.7500
## Pos Pred Value : 0.6154
## Neg Pred Value : 0.5882
## Prevalence : 0.4805
## Detection Rate : 0.2078
## Detection Prevalence : 0.3377
## Balanced Accuracy : 0.5912
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_3B.2_pred <- predict(smile__svm_model_3B, tst_smile_girls)
summary(smile__svm_model_3B.2_pred)
## spontaneous deliberate
## 24 41
smile__svm_model_3B.2_confM <- confusionMatrix(
smile__svm_model_3B.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_3B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 14 10
## deliberate 19 22
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.2678
##
## Kappa : 0.1113
##
## Mcnemar's Test P-Value : 0.1374
##
## Sensitivity : 0.4242
## Specificity : 0.6875
## Pos Pred Value : 0.5833
## Neg Pred Value : 0.5366
## Prevalence : 0.5077
## Detection Rate : 0.2154
## Detection Prevalence : 0.3692
## Balanced Accuracy : 0.5559
##
## 'Positive' Class : spontaneous
##
# model 3C lip
set.seed(1973)
smile__svm_model_3C <- train(smile_type ~ eye_mean + lip_mean + amplitude_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_3C$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5143661 0.02580907 0.08044698 0.1609386
summary(smile__svm_model_3C$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_3C_pred <- predict(smile__svm_model_3C, tst_smile)
summary(smile__svm_model_3C_pred)
## spontaneous deliberate
## 55 87
smile__svm_model_3C_confM <- confusionMatrix(
smile__svm_model_3C_pred,
tst_smile$smile_type
)
smile__svm_model_3C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 31 24
## deliberate 39 48
##
## Accuracy : 0.5563
## 95% CI : (0.4707, 0.6396)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.13758
##
## Kappa : 0.1099
##
## Mcnemar's Test P-Value : 0.07776
##
## Sensitivity : 0.4429
## Specificity : 0.6667
## Pos Pred Value : 0.5636
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.2183
## Detection Prevalence : 0.3873
## Balanced Accuracy : 0.5548
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_3C.1_pred <- predict(smile__svm_model_3C, tst_smile_boys)
summary(smile__svm_model_3C.1_pred)
## spontaneous deliberate
## 28 49
smile__svm_model_3C.1_confM <- confusionMatrix(
smile__svm_model_3C.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_3C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 17 11
## deliberate 20 29
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.1862
##
## Mcnemar's Test P-Value : 0.1508
##
## Sensitivity : 0.4595
## Specificity : 0.7250
## Pos Pred Value : 0.6071
## Neg Pred Value : 0.5918
## Prevalence : 0.4805
## Detection Rate : 0.2208
## Detection Prevalence : 0.3636
## Balanced Accuracy : 0.5922
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_3C.2_pred <- predict(smile__svm_model_3C, tst_smile_girls)
summary(smile__svm_model_3C.2_pred)
## spontaneous deliberate
## 27 38
smile__svm_model_3C.2_confM <- confusionMatrix(
smile__svm_model_3C.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_3C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 14 13
## deliberate 19 19
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0179
##
## Mcnemar's Test P-Value : 0.3768
##
## Sensitivity : 0.4242
## Specificity : 0.5938
## Pos Pred Value : 0.5185
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.2154
## Detection Prevalence : 0.4154
## Balanced Accuracy : 0.5090
##
## 'Positive' Class : spontaneous
##
# model 4 AU features
set.seed(1973)
smile__svm_model_4 <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6932765 0.3879683 0.1209849 0.2395002
summary(smile__svm_model_4$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4_pred <- predict(smile__svm_model_4, tst_smile)
summary(smile__svm_model_4_pred)
## spontaneous deliberate
## 60 82
smile__svm_model_4_confM <- confusionMatrix(
smile__svm_model_4_pred,
tst_smile$smile_type
)
smile__svm_model_4_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 35 25
## deliberate 35 47
##
## Accuracy : 0.5775
## 95% CI : (0.4918, 0.6598)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.05517
##
## Kappa : 0.1531
##
## Mcnemar's Test P-Value : 0.24528
##
## Sensitivity : 0.5000
## Specificity : 0.6528
## Pos Pred Value : 0.5833
## Neg Pred Value : 0.5732
## Prevalence : 0.4930
## Detection Rate : 0.2465
## Detection Prevalence : 0.4225
## Balanced Accuracy : 0.5764
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4.1_pred <- predict(smile__svm_model_4, tst_smile_boys)
summary(smile__svm_model_4.1_pred)
## spontaneous deliberate
## 34 43
smile__svm_model_4.1_confM <- confusionMatrix(
smile__svm_model_4.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 13
## deliberate 16 27
##
## Accuracy : 0.6234
## 95% CI : (0.5056, 0.7313)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.04297
##
## Kappa : 0.2433
##
## Mcnemar's Test P-Value : 0.71035
##
## Sensitivity : 0.5676
## Specificity : 0.6750
## Pos Pred Value : 0.6176
## Neg Pred Value : 0.6279
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.4416
## Balanced Accuracy : 0.6213
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4.2_pred <- predict(smile__svm_model_4, tst_smile_girls)
summary(smile__svm_model_4.2_pred)
## spontaneous deliberate
## 26 39
smile__svm_model_4.2_confM <- confusionMatrix(
smile__svm_model_4.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 14 12
## deliberate 19 20
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.0491
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.4242
## Specificity : 0.6250
## Pos Pred Value : 0.5385
## Neg Pred Value : 0.5128
## Prevalence : 0.5077
## Detection Rate : 0.2154
## Detection Prevalence : 0.4000
## Balanced Accuracy : 0.5246
##
## 'Positive' Class : spontaneous
##
# model 4A AU happiness model
set.seed(1973)
smile__svm_model_4A <- train(smile_type ~ AU06_r_mean + AU12_r_mean,
method = "svmLinear",
data = trn_smile,
trControl = trainControl(
method = "cv",
number = 10
)
)
smile__svm_model_4A$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5249387 0.04894662 0.07451885 0.1516131
summary(smile__svm_model_4A$finalModel)
## Length Class Mode
## 1 ksvm S4
kernlab::plot(smile__svm_model_4A$finalModel, xlab = "AU12", ylab = "AU06")
smile__svm_model_4A_pred <- predict(smile__svm_model_4A, tst_smile)
summary(smile__svm_model_4A_pred)
## spontaneous deliberate
## 41 101
smile__svm_model_4A_confM <- confusionMatrix(
smile__svm_model_4A_pred,
tst_smile$smile_type
)
smile__svm_model_4A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 18
## deliberate 47 54
##
## Accuracy : 0.5423
## 95% CI : (0.4567, 0.6261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2251326
##
## Kappa : 0.079
##
## Mcnemar's Test P-Value : 0.0005147
##
## Sensitivity : 0.3286
## Specificity : 0.7500
## Pos Pred Value : 0.5610
## Neg Pred Value : 0.5347
## Prevalence : 0.4930
## Detection Rate : 0.1620
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.5393
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4A.1_pred <- predict(smile__svm_model_4A, tst_smile_boys)
summary(smile__svm_model_4A.1_pred)
## spontaneous deliberate
## 17 60
smile__svm_model_4A.1_confM <- confusionMatrix(
smile__svm_model_4A.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 10 7
## deliberate 27 33
##
## Accuracy : 0.5584
## 95% CI : (0.4407, 0.6716)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.28475
##
## Kappa : 0.0972
##
## Mcnemar's Test P-Value : 0.00112
##
## Sensitivity : 0.2703
## Specificity : 0.8250
## Pos Pred Value : 0.5882
## Neg Pred Value : 0.5500
## Prevalence : 0.4805
## Detection Rate : 0.1299
## Detection Prevalence : 0.2208
## Balanced Accuracy : 0.5476
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4A.2_pred <- predict(smile__svm_model_4A, tst_smile_girls)
summary(smile__svm_model_4A.2_pred)
## spontaneous deliberate
## 24 41
smile__svm_model_4A.2_confM <- confusionMatrix(
smile__svm_model_4A.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 11
## deliberate 20 21
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.4510
##
## Kappa : 0.05
##
## Mcnemar's Test P-Value : 0.1508
##
## Sensitivity : 0.3939
## Specificity : 0.6562
## Pos Pred Value : 0.5417
## Neg Pred Value : 0.5122
## Prevalence : 0.5077
## Detection Rate : 0.2000
## Detection Prevalence : 0.3692
## Balanced Accuracy : 0.5251
##
## 'Positive' Class : spontaneous
##
# model 4B AU best model
set.seed(1973)
smile__svm_model_4B <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean,
method = "svmLinear",
data = trn_smile,
trControl = trainControl(
method = "cv",
number = 10
)
)
smile__svm_model_4B$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6304089 0.2607284 0.05916065 0.1174346
summary(smile__svm_model_4B$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4B_pred <- predict(smile__svm_model_4B, tst_smile)
summary(smile__svm_model_4B_pred)
## spontaneous deliberate
## 63 79
smile__svm_model_4B_confM <- confusionMatrix(
smile__svm_model_4B_pred,
tst_smile$smile_type
)
smile__svm_model_4B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 41 22
## deliberate 29 50
##
## Accuracy : 0.6408
## 95% CI : (0.5561, 0.7196)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0008891
##
## Kappa : 0.2805
##
## Mcnemar's Test P-Value : 0.4008142
##
## Sensitivity : 0.5857
## Specificity : 0.6944
## Pos Pred Value : 0.6508
## Neg Pred Value : 0.6329
## Prevalence : 0.4930
## Detection Rate : 0.2887
## Detection Prevalence : 0.4437
## Balanced Accuracy : 0.6401
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4B.1_pred <- predict(smile__svm_model_4B, tst_smile_boys)
summary(smile__svm_model_4B.1_pred)
## spontaneous deliberate
## 26 51
smile__svm_model_4B.1_confM <- confusionMatrix(
smile__svm_model_4B.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 6
## deliberate 17 34
##
## Accuracy : 0.7013
## 95% CI : (0.5862, 0.8003)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.0008966
##
## Kappa : 0.3949
##
## Mcnemar's Test P-Value : 0.0370562
##
## Sensitivity : 0.5405
## Specificity : 0.8500
## Pos Pred Value : 0.7692
## Neg Pred Value : 0.6667
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.3377
## Balanced Accuracy : 0.6953
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4B.2_pred <- predict(smile__svm_model_4B, tst_smile_girls)
summary(smile__svm_model_4B.2_pred)
## spontaneous deliberate
## 37 28
smile__svm_model_4B.2_confM <- confusionMatrix(
smile__svm_model_4B.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 16
## deliberate 12 16
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.1927
##
## Kappa : 0.1366
##
## Mcnemar's Test P-Value : 0.5708
##
## Sensitivity : 0.6364
## Specificity : 0.5000
## Pos Pred Value : 0.5676
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.5692
## Balanced Accuracy : 0.5682
##
## 'Positive' Class : spontaneous
##
# model 4C AU happiness model + blink
set.seed(1973)
smile__svm_model_4C <- train(smile_type ~ AU45_r_mean + AU06_r_mean +
AU12_r_mean,
method = "svmLinear",
data = trn_smile,
trControl = trainControl(
method = "cv",
number = 10
)
)
smile__svm_model_4C$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5854947 0.1737161 0.0594244 0.1205028
summary(smile__svm_model_4C$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4C_pred <- predict(smile__svm_model_4C, tst_smile)
summary(smile__svm_model_4C_pred)
## spontaneous deliberate
## 99 43
smile__svm_model_4C_confM <- confusionMatrix(
smile__svm_model_4C_pred,
tst_smile$smile_type
)
smile__svm_model_4C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 52 47
## deliberate 18 25
##
## Accuracy : 0.5423
## 95% CI : (0.4567, 0.6261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2251326
##
## Kappa : 0.0896
##
## Mcnemar's Test P-Value : 0.0005147
##
## Sensitivity : 0.7429
## Specificity : 0.3472
## Pos Pred Value : 0.5253
## Neg Pred Value : 0.5814
## Prevalence : 0.4930
## Detection Rate : 0.3662
## Detection Prevalence : 0.6972
## Balanced Accuracy : 0.5450
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4C.1_pred <- predict(smile__svm_model_4C, tst_smile_boys)
summary(smile__svm_model_4C.1_pred)
## spontaneous deliberate
## 50 27
smile__svm_model_4C.1_confM <- confusionMatrix(
smile__svm_model_4C.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 24
## deliberate 11 16
##
## Accuracy : 0.5455
## 95% CI : (0.4279, 0.6594)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.36674
##
## Kappa : 0.1014
##
## Mcnemar's Test P-Value : 0.04252
##
## Sensitivity : 0.7027
## Specificity : 0.4000
## Pos Pred Value : 0.5200
## Neg Pred Value : 0.5926
## Prevalence : 0.4805
## Detection Rate : 0.3377
## Detection Prevalence : 0.6494
## Balanced Accuracy : 0.5514
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4C.2_pred <- predict(smile__svm_model_4C, tst_smile_girls)
summary(smile__svm_model_4C.2_pred)
## spontaneous deliberate
## 49 16
smile__svm_model_4C.2_confM <- confusionMatrix(
smile__svm_model_4C.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 23
## deliberate 7 9
##
## Accuracy : 0.5385
## 95% CI : (0.4103, 0.663)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.35525
##
## Kappa : 0.0697
##
## Mcnemar's Test P-Value : 0.00617
##
## Sensitivity : 0.7879
## Specificity : 0.2812
## Pos Pred Value : 0.5306
## Neg Pred Value : 0.5625
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.7538
## Balanced Accuracy : 0.5346
##
## 'Positive' Class : spontaneous
##
# model 4D AU45
set.seed(1973)
smile__svm_model_4D <-
train(smile_type ~ AU45_r_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4D$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5853164 0.1741181 0.05423172 0.1090408
summary(smile__svm_model_4D$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4D_pred <- predict(smile__svm_model_4D, tst_smile)
summary(smile__svm_model_4D_pred)
## spontaneous deliberate
## 96 46
smile__svm_model_4D_confM <- confusionMatrix(
smile__svm_model_4D_pred,
tst_smile$smile_type
)
smile__svm_model_4D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 49 47
## deliberate 21 25
##
## Accuracy : 0.5211
## 95% CI : (0.4358, 0.6056)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.400803
##
## Kappa : 0.047
##
## Mcnemar's Test P-Value : 0.002432
##
## Sensitivity : 0.7000
## Specificity : 0.3472
## Pos Pred Value : 0.5104
## Neg Pred Value : 0.5435
## Prevalence : 0.4930
## Detection Rate : 0.3451
## Detection Prevalence : 0.6761
## Balanced Accuracy : 0.5236
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4D.1_pred <- predict(smile__svm_model_4D, tst_smile_boys)
summary(smile__svm_model_4D.1_pred)
## spontaneous deliberate
## 51 26
smile__svm_model_4D.1_confM <- confusionMatrix(
smile__svm_model_4D.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 26
## deliberate 12 14
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.63425
##
## Kappa : 0.0253
##
## Mcnemar's Test P-Value : 0.03496
##
## Sensitivity : 0.6757
## Specificity : 0.3500
## Pos Pred Value : 0.4902
## Neg Pred Value : 0.5385
## Prevalence : 0.4805
## Detection Rate : 0.3247
## Detection Prevalence : 0.6623
## Balanced Accuracy : 0.5128
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4D.2_pred <- predict(smile__svm_model_4D, tst_smile_girls)
summary(smile__svm_model_4D.2_pred)
## spontaneous deliberate
## 45 20
smile__svm_model_4D.2_confM <- confusionMatrix(
smile__svm_model_4D.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 21
## deliberate 9 11
##
## Accuracy : 0.5385
## 95% CI : (0.4103, 0.663)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.35525
##
## Kappa : 0.0714
##
## Mcnemar's Test P-Value : 0.04461
##
## Sensitivity : 0.7273
## Specificity : 0.3438
## Pos Pred Value : 0.5333
## Neg Pred Value : 0.5500
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.6923
## Balanced Accuracy : 0.5355
##
## 'Positive' Class : spontaneous
##
# model 4E AU12
set.seed(1973)
smile__svm_model_4E <-
train(smile_type ~ AU12_r_mean,
method = "svmLinear",
data = trn_smile, trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4E$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5399287 0.08094986 0.05983632 0.1207807
summary(smile__svm_model_4E$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4E_pred <- predict(smile__svm_model_4E, tst_smile)
summary(smile__svm_model_4E_pred)
## spontaneous deliberate
## 55 87
smile__svm_model_4E_confM <- confusionMatrix(
smile__svm_model_4E_pred,
tst_smile$smile_type
)
smile__svm_model_4E_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 28 27
## deliberate 42 45
##
## Accuracy : 0.5141
## 95% CI : (0.4288, 0.5987)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.46673
##
## Kappa : 0.0251
##
## Mcnemar's Test P-Value : 0.09191
##
## Sensitivity : 0.4000
## Specificity : 0.6250
## Pos Pred Value : 0.5091
## Neg Pred Value : 0.5172
## Prevalence : 0.4930
## Detection Rate : 0.1972
## Detection Prevalence : 0.3873
## Balanced Accuracy : 0.5125
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4E.1_pred <- predict(smile__svm_model_4E, tst_smile_boys)
summary(smile__svm_model_4E.1_pred)
## spontaneous deliberate
## 20 57
smile__svm_model_4E.1_confM <- confusionMatrix(
smile__svm_model_4E.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 8
## deliberate 25 32
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.212593
##
## Kappa : 0.1265
##
## Mcnemar's Test P-Value : 0.005349
##
## Sensitivity : 0.3243
## Specificity : 0.8000
## Pos Pred Value : 0.6000
## Neg Pred Value : 0.5614
## Prevalence : 0.4805
## Detection Rate : 0.1558
## Detection Prevalence : 0.2597
## Balanced Accuracy : 0.5622
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4E.2_pred <- predict(smile__svm_model_4E, tst_smile_girls)
summary(smile__svm_model_4E.2_pred)
## spontaneous deliberate
## 35 30
smile__svm_model_4E.2_confM <- confusionMatrix(
smile__svm_model_4E.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 19
## deliberate 17 13
##
## Accuracy : 0.4462
## 95% CI : (0.3227, 0.5747)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.8679
##
## Kappa : -0.109
##
## Mcnemar's Test P-Value : 0.8676
##
## Sensitivity : 0.4848
## Specificity : 0.4062
## Pos Pred Value : 0.4571
## Neg Pred Value : 0.4333
## Prevalence : 0.5077
## Detection Rate : 0.2462
## Detection Prevalence : 0.5385
## Balanced Accuracy : 0.4455
##
## 'Positive' Class : spontaneous
##
# model 4F AU06
set.seed(1973)
smile__svm_model_4F <-
train(smile_type ~ AU06_r_mean,
method = "svmLinear",
data = trn_smile, trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4F$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5487411 0.09600249 0.07076652 0.1449534
summary(smile__svm_model_4F$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4F_pred <- predict(smile__svm_model_4F, tst_smile)
summary(smile__svm_model_4F_pred)
## spontaneous deliberate
## 44 98
smile__svm_model_4F_confM <- confusionMatrix(
smile__svm_model_4F_pred,
tst_smile$smile_type
)
smile__svm_model_4F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 20
## deliberate 46 52
##
## Accuracy : 0.5352
## 95% CI : (0.4497, 0.6193)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.278603
##
## Kappa : 0.0654
##
## Mcnemar's Test P-Value : 0.002089
##
## Sensitivity : 0.3429
## Specificity : 0.7222
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5306
## Prevalence : 0.4930
## Detection Rate : 0.1690
## Detection Prevalence : 0.3099
## Balanced Accuracy : 0.5325
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4F.1_pred <- predict(smile__svm_model_4F, tst_smile_boys)
summary(smile__svm_model_4F.1_pred)
## spontaneous deliberate
## 17 60
smile__svm_model_4F.1_confM <- confusionMatrix(
smile__svm_model_4F.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 10 7
## deliberate 27 33
##
## Accuracy : 0.5584
## 95% CI : (0.4407, 0.6716)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.28475
##
## Kappa : 0.0972
##
## Mcnemar's Test P-Value : 0.00112
##
## Sensitivity : 0.2703
## Specificity : 0.8250
## Pos Pred Value : 0.5882
## Neg Pred Value : 0.5500
## Prevalence : 0.4805
## Detection Rate : 0.1299
## Detection Prevalence : 0.2208
## Balanced Accuracy : 0.5476
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4F.2_pred <- predict(smile__svm_model_4F, tst_smile_girls)
summary(smile__svm_model_4F.2_pred)
## spontaneous deliberate
## 27 38
smile__svm_model_4F.2_confM <- confusionMatrix(
smile__svm_model_4F.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 14 13
## deliberate 19 19
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0179
##
## Mcnemar's Test P-Value : 0.3768
##
## Sensitivity : 0.4242
## Specificity : 0.5938
## Pos Pred Value : 0.5185
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.2154
## Detection Prevalence : 0.4154
## Balanced Accuracy : 0.5090
##
## 'Positive' Class : spontaneous
##
# model 4G AU10
set.seed(1973)
smile__svm_model_4G <-
train(smile_type ~ AU10_r_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4G$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5729 0.1451991 0.08629806 0.1728594
summary(smile__svm_model_4G$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4G_pred <- predict(smile__svm_model_4G, tst_smile)
summary(smile__svm_model_4G_pred)
## spontaneous deliberate
## 52 90
smile__svm_model_4G_confM <- confusionMatrix(
smile__svm_model_4G_pred,
tst_smile$smile_type
)
smile__svm_model_4G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 29 23
## deliberate 41 49
##
## Accuracy : 0.5493
## 95% CI : (0.4636, 0.6328)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.17799
##
## Kappa : 0.0952
##
## Mcnemar's Test P-Value : 0.03359
##
## Sensitivity : 0.4143
## Specificity : 0.6806
## Pos Pred Value : 0.5577
## Neg Pred Value : 0.5444
## Prevalence : 0.4930
## Detection Rate : 0.2042
## Detection Prevalence : 0.3662
## Balanced Accuracy : 0.5474
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4G.1_pred <- predict(smile__svm_model_4G, tst_smile_boys)
summary(smile__svm_model_4G.1_pred)
## spontaneous deliberate
## 19 58
smile__svm_model_4G.1_confM <- confusionMatrix(
smile__svm_model_4G.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 13 6
## deliberate 24 34
##
## Accuracy : 0.6104
## 95% CI : (0.4925, 0.7195)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.068593
##
## Kappa : 0.2051
##
## Mcnemar's Test P-Value : 0.001911
##
## Sensitivity : 0.3514
## Specificity : 0.8500
## Pos Pred Value : 0.6842
## Neg Pred Value : 0.5862
## Prevalence : 0.4805
## Detection Rate : 0.1688
## Detection Prevalence : 0.2468
## Balanced Accuracy : 0.6007
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4G.2_pred <- predict(smile__svm_model_4G, tst_smile_girls)
summary(smile__svm_model_4G.2_pred)
## spontaneous deliberate
## 33 32
smile__svm_model_4G.2_confM <- confusionMatrix(
smile__svm_model_4G.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 17
## deliberate 17 15
##
## Accuracy : 0.4769
## 95% CI : (0.3515, 0.6046)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.7324
##
## Kappa : -0.0464
##
## Mcnemar's Test P-Value : 1.0000
##
## Sensitivity : 0.4848
## Specificity : 0.4688
## Pos Pred Value : 0.4848
## Neg Pred Value : 0.4688
## Prevalence : 0.5077
## Detection Rate : 0.2462
## Detection Prevalence : 0.5077
## Balanced Accuracy : 0.4768
##
## 'Positive' Class : spontaneous
##
# model 4H AU01
set.seed(1973)
smile__svm_model_4H <-
train(smile_type ~ AU01_r_mean,
method = "svmLinear",
data = trn_smile, trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4H$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5350546 0.07469541 0.08066542 0.1592671
summary(smile__svm_model_4H$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4H_pred <- predict(smile__svm_model_4H, tst_smile)
summary(smile__svm_model_4H_pred)
## spontaneous deliberate
## 115 27
smile__svm_model_4H_confM <- confusionMatrix(
smile__svm_model_4H_pred,
tst_smile$smile_type
)
smile__svm_model_4H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 60 55
## deliberate 10 17
##
## Accuracy : 0.5423
## 95% CI : (0.4567, 0.6261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.2251
##
## Kappa : 0.0924
##
## Mcnemar's Test P-Value : 4.828e-08
##
## Sensitivity : 0.8571
## Specificity : 0.2361
## Pos Pred Value : 0.5217
## Neg Pred Value : 0.6296
## Prevalence : 0.4930
## Detection Rate : 0.4225
## Detection Prevalence : 0.8099
## Balanced Accuracy : 0.5466
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4H.1_pred <- predict(smile__svm_model_4H, tst_smile_boys)
summary(smile__svm_model_4H.1_pred)
## spontaneous deliberate
## 65 12
smile__svm_model_4H.1_confM <- confusionMatrix(
smile__svm_model_4H.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 34 31
## deliberate 3 9
##
## Accuracy : 0.5584
## 95% CI : (0.4407, 0.6716)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.2847
##
## Kappa : 0.1399
##
## Mcnemar's Test P-Value : 3.649e-06
##
## Sensitivity : 0.9189
## Specificity : 0.2250
## Pos Pred Value : 0.5231
## Neg Pred Value : 0.7500
## Prevalence : 0.4805
## Detection Rate : 0.4416
## Detection Prevalence : 0.8442
## Balanced Accuracy : 0.5720
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4H.2_pred <- predict(smile__svm_model_4H, tst_smile_girls)
summary(smile__svm_model_4H.2_pred)
## spontaneous deliberate
## 50 15
smile__svm_model_4H.2_confM <- confusionMatrix(
smile__svm_model_4H.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 24
## deliberate 7 8
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.450951
##
## Kappa : 0.0382
##
## Mcnemar's Test P-Value : 0.004057
##
## Sensitivity : 0.7879
## Specificity : 0.2500
## Pos Pred Value : 0.5200
## Neg Pred Value : 0.5333
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.7692
## Balanced Accuracy : 0.5189
##
## 'Positive' Class : spontaneous
##
# model 4I AU25
set.seed(1973)
smile__svm_model_4I <-
train(smile_type ~ AU25_r_mean,
method = "svmLinear",
data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4I$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5705158 0.143362 0.04911462 0.09427582
summary(smile__svm_model_4I$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4I_pred <- predict(smile__svm_model_4I, tst_smile)
summary(smile__svm_model_4I_pred)
## spontaneous deliberate
## 106 36
smile__svm_model_4I_confM <- confusionMatrix(
smile__svm_model_4I_pred,
tst_smile$smile_type
)
smile__svm_model_4I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 60 46
## deliberate 10 26
##
## Accuracy : 0.6056
## 95% CI : (0.5202, 0.6865)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.01151
##
## Kappa : 0.2167
##
## Mcnemar's Test P-Value : 2.91e-06
##
## Sensitivity : 0.8571
## Specificity : 0.3611
## Pos Pred Value : 0.5660
## Neg Pred Value : 0.7222
## Prevalence : 0.4930
## Detection Rate : 0.4225
## Detection Prevalence : 0.7465
## Balanced Accuracy : 0.6091
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4I.1_pred <- predict(smile__svm_model_4I, tst_smile_boys)
summary(smile__svm_model_4I.1_pred)
## spontaneous deliberate
## 62 15
smile__svm_model_4I.1_confM <- confusionMatrix(
smile__svm_model_4I.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 34 28
## deliberate 3 12
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.2135
##
## Mcnemar's Test P-Value : 1.629e-05
##
## Sensitivity : 0.9189
## Specificity : 0.3000
## Pos Pred Value : 0.5484
## Neg Pred Value : 0.8000
## Prevalence : 0.4805
## Detection Rate : 0.4416
## Detection Prevalence : 0.8052
## Balanced Accuracy : 0.6095
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4I.2_pred <- predict(smile__svm_model_4I, tst_smile_girls)
summary(smile__svm_model_4I.2_pred)
## spontaneous deliberate
## 44 21
smile__svm_model_4I.2_confM <- confusionMatrix(
smile__svm_model_4I.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 18
## deliberate 7 14
##
## Accuracy : 0.6154
## 95% CI : (0.4864, 0.7335)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.05294
##
## Kappa : 0.2266
##
## Mcnemar's Test P-Value : 0.04550
##
## Sensitivity : 0.7879
## Specificity : 0.4375
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.6127
##
## 'Positive' Class : spontaneous
##
# model 4J AU09
set.seed(1973)
smile__svm_model_4J <-
train(smile_type ~ AU09_r_mean,
method = "svmLinear",
data = trn_smile, trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_4J$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5249276 0.04510716 0.08572585 0.1669863
summary(smile__svm_model_4J$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_4J_pred <- predict(smile__svm_model_4J, tst_smile)
summary(smile__svm_model_4J_pred)
## spontaneous deliberate
## 30 112
smile__svm_model_4J_confM <- confusionMatrix(
smile__svm_model_4J_pred,
tst_smile$smile_type
)
smile__svm_model_4J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 18
## deliberate 58 54
##
## Accuracy : 0.4648
## 95% CI : (0.3807, 0.5503)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.8624
##
## Kappa : -0.0792
##
## Mcnemar's Test P-Value : 7.691e-06
##
## Sensitivity : 0.17143
## Specificity : 0.75000
## Pos Pred Value : 0.40000
## Neg Pred Value : 0.48214
## Prevalence : 0.49296
## Detection Rate : 0.08451
## Detection Prevalence : 0.21127
## Balanced Accuracy : 0.46071
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_4J.1_pred <- predict(smile__svm_model_4J, tst_smile_boys)
summary(smile__svm_model_4J.1_pred)
## spontaneous deliberate
## 16 61
smile__svm_model_4J.1_confM <- confusionMatrix(
smile__svm_model_4J.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_4J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 6 10
## deliberate 31 30
##
## Accuracy : 0.4675
## 95% CI : (0.3529, 0.5848)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.847647
##
## Kappa : -0.0897
##
## Mcnemar's Test P-Value : 0.001787
##
## Sensitivity : 0.16216
## Specificity : 0.75000
## Pos Pred Value : 0.37500
## Neg Pred Value : 0.49180
## Prevalence : 0.48052
## Detection Rate : 0.07792
## Detection Prevalence : 0.20779
## Balanced Accuracy : 0.45608
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_4J.2_pred <- predict(smile__svm_model_4J, tst_smile_girls)
summary(smile__svm_model_4J.2_pred)
## spontaneous deliberate
## 14 51
smile__svm_model_4J.2_confM <- confusionMatrix(
smile__svm_model_4J.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_4J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 6 8
## deliberate 27 24
##
## Accuracy : 0.4615
## 95% CI : (0.337, 0.5897)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.807360
##
## Kappa : -0.0676
##
## Mcnemar's Test P-Value : 0.002346
##
## Sensitivity : 0.18182
## Specificity : 0.75000
## Pos Pred Value : 0.42857
## Neg Pred Value : 0.47059
## Prevalence : 0.50769
## Detection Rate : 0.09231
## Detection Prevalence : 0.21538
## Balanced Accuracy : 0.46591
##
## 'Positive' Class : spontaneous
##
# model 5 head pose features
set.seed(1973)
smile__svm_model_5 <-
train(smile_type ~ pose_Rx_mean + pose_Ry_mean + pose_Rz_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_5$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.4505236 -0.09969948 0.06191438 0.1240377
summary(smile__svm_model_5$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_5_pred <- predict(smile__svm_model_5, tst_smile)
summary(smile__svm_model_5_pred)
## spontaneous deliberate
## 62 80
smile__svm_model_5_confM <- confusionMatrix(
smile__svm_model_5_pred,
tst_smile$smile_type
)
smile__svm_model_5_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 31 31
## deliberate 39 41
##
## Accuracy : 0.507
## 95% CI : (0.4219, 0.5919)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.5336
##
## Kappa : 0.0123
##
## Mcnemar's Test P-Value : 0.4028
##
## Sensitivity : 0.4429
## Specificity : 0.5694
## Pos Pred Value : 0.5000
## Neg Pred Value : 0.5125
## Prevalence : 0.4930
## Detection Rate : 0.2183
## Detection Prevalence : 0.4366
## Balanced Accuracy : 0.5062
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_5.1_pred <- predict(smile__svm_model_5, tst_smile_boys)
summary(smile__svm_model_5.1_pred)
## spontaneous deliberate
## 31 46
smile__svm_model_5.1_confM <- confusionMatrix(
smile__svm_model_5.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_5.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 15 16
## deliberate 22 24
##
## Accuracy : 0.5065
## 95% CI : (0.39, 0.6224)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.6342
##
## Kappa : 0.0054
##
## Mcnemar's Test P-Value : 0.4173
##
## Sensitivity : 0.4054
## Specificity : 0.6000
## Pos Pred Value : 0.4839
## Neg Pred Value : 0.5217
## Prevalence : 0.4805
## Detection Rate : 0.1948
## Detection Prevalence : 0.4026
## Balanced Accuracy : 0.5027
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_5.2_pred <- predict(smile__svm_model_5, tst_smile_girls)
summary(smile__svm_model_5.2_pred)
## spontaneous deliberate
## 31 34
smile__svm_model_5.2_confM <- confusionMatrix(
smile__svm_model_5.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_5.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 15
## deliberate 17 17
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0161
##
## Mcnemar's Test P-Value : 0.8597
##
## Sensitivity : 0.4848
## Specificity : 0.5312
## Pos Pred Value : 0.5161
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.2462
## Detection Prevalence : 0.4769
## Balanced Accuracy : 0.5080
##
## 'Positive' Class : spontaneous
##
# model 5A gaze features
set.seed(1973)
smile__svm_model_5A <- train(smile_type ~ gaze_angle_x_mean + gaze_angle_y_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_5A$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.4712901 -0.05969345 0.04430412 0.08370863
summary(smile__svm_model_5A$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_5A_pred <- predict(smile__svm_model_5A, tst_smile)
summary(smile__svm_model_5A_pred)
## spontaneous deliberate
## 41 101
smile__svm_model_5A_confM <- confusionMatrix(
smile__svm_model_5A_pred,
tst_smile$smile_type
)
smile__svm_model_5A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 30 11
## deliberate 40 61
##
## Accuracy : 0.6408
## 95% CI : (0.5561, 0.7196)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0008891
##
## Kappa : 0.2774
##
## Mcnemar's Test P-Value : 8.826e-05
##
## Sensitivity : 0.4286
## Specificity : 0.8472
## Pos Pred Value : 0.7317
## Neg Pred Value : 0.6040
## Prevalence : 0.4930
## Detection Rate : 0.2113
## Detection Prevalence : 0.2887
## Balanced Accuracy : 0.6379
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_5A.1_pred <- predict(smile__svm_model_5A, tst_smile_boys)
summary(smile__svm_model_5A.1_pred)
## spontaneous deliberate
## 27 50
smile__svm_model_5A.1_confM <- confusionMatrix(
smile__svm_model_5A.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_5A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 18 9
## deliberate 19 31
##
## Accuracy : 0.6364
## 95% CI : (0.5188, 0.743)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.02565
##
## Kappa : 0.2642
##
## Mcnemar's Test P-Value : 0.08897
##
## Sensitivity : 0.4865
## Specificity : 0.7750
## Pos Pred Value : 0.6667
## Neg Pred Value : 0.6200
## Prevalence : 0.4805
## Detection Rate : 0.2338
## Detection Prevalence : 0.3506
## Balanced Accuracy : 0.6307
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_5A.2_pred <- predict(smile__svm_model_5A, tst_smile_girls)
summary(smile__svm_model_5A.2_pred)
## spontaneous deliberate
## 14 51
smile__svm_model_5A.2_confM <- confusionMatrix(
smile__svm_model_5A.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_5A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 12 2
## deliberate 21 30
##
## Accuracy : 0.6462
## 95% CI : (0.5177, 0.7608)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.0169758
##
## Kappa : 0.2985
##
## Mcnemar's Test P-Value : 0.0001746
##
## Sensitivity : 0.3636
## Specificity : 0.9375
## Pos Pred Value : 0.8571
## Neg Pred Value : 0.5882
## Prevalence : 0.5077
## Detection Rate : 0.1846
## Detection Prevalence : 0.2154
## Balanced Accuracy : 0.6506
##
## 'Positive' Class : spontaneous
##
# model 5B gaze + head pose features
set.seed(1973)
smile__svm_model_5B <-
train(smile_type ~ pose_Rx_mean + pose_Ry_mean + pose_Rz_mean +
gaze_angle_x_mean + gaze_angle_y_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_5B$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.4737745 -0.05205384 0.0817408 0.162664
summary(smile__svm_model_5B$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_5B_pred <- predict(smile__svm_model_5B, tst_smile)
summary(smile__svm_model_5B_pred)
## spontaneous deliberate
## 55 87
smile__svm_model_5B_confM <- confusionMatrix(
smile__svm_model_5B_pred,
tst_smile$smile_type
)
smile__svm_model_5B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 31 24
## deliberate 39 48
##
## Accuracy : 0.5563
## 95% CI : (0.4707, 0.6396)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.13758
##
## Kappa : 0.1099
##
## Mcnemar's Test P-Value : 0.07776
##
## Sensitivity : 0.4429
## Specificity : 0.6667
## Pos Pred Value : 0.5636
## Neg Pred Value : 0.5517
## Prevalence : 0.4930
## Detection Rate : 0.2183
## Detection Prevalence : 0.3873
## Balanced Accuracy : 0.5548
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_5B.1_pred <- predict(smile__svm_model_5B, tst_smile_boys)
summary(smile__svm_model_5B.1_pred)
## spontaneous deliberate
## 30 47
smile__svm_model_5B.1_confM <- confusionMatrix(
smile__svm_model_5B.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_5B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 14 16
## deliberate 23 24
##
## Accuracy : 0.4935
## 95% CI : (0.3776, 0.61)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.7159
##
## Kappa : -0.0218
##
## Mcnemar's Test P-Value : 0.3367
##
## Sensitivity : 0.3784
## Specificity : 0.6000
## Pos Pred Value : 0.4667
## Neg Pred Value : 0.5106
## Prevalence : 0.4805
## Detection Rate : 0.1818
## Detection Prevalence : 0.3896
## Balanced Accuracy : 0.4892
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_5B.2_pred <- predict(smile__svm_model_5B, tst_smile_girls)
summary(smile__svm_model_5B.2_pred)
## spontaneous deliberate
## 25 40
smile__svm_model_5B.2_confM <- confusionMatrix(
smile__svm_model_5B.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_5B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 17 8
## deliberate 16 24
##
## Accuracy : 0.6308
## 95% CI : (0.502, 0.7472)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.03086
##
## Kappa : 0.2642
##
## Mcnemar's Test P-Value : 0.15304
##
## Sensitivity : 0.5152
## Specificity : 0.7500
## Pos Pred Value : 0.6800
## Neg Pred Value : 0.6000
## Prevalence : 0.5077
## Detection Rate : 0.2615
## Detection Prevalence : 0.3846
## Balanced Accuracy : 0.6326
##
## 'Positive' Class : spontaneous
##
# model 6 dynamics and movement
set.seed(1973)
smile__svm_model_6 <-
train(smile_type ~ onset_mean + apex_mean + offset_mean + eye_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.665781 0.3333057 0.08962779 0.1773316
summary(smile__svm_model_6$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_6_pred <- predict(smile__svm_model_6, tst_smile)
summary(smile__svm_model_6_pred)
## spontaneous deliberate
## 74 68
smile__svm_model_6_confM <- confusionMatrix(
smile__svm_model_6_pred,
tst_smile$smile_type
)
smile__svm_model_6_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 48 26
## deliberate 22 46
##
## Accuracy : 0.662
## 95% CI : (0.5779, 0.7391)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0001362
##
## Kappa : 0.3243
##
## Mcnemar's Test P-Value : 0.6650055
##
## Sensitivity : 0.6857
## Specificity : 0.6389
## Pos Pred Value : 0.6486
## Neg Pred Value : 0.6765
## Prevalence : 0.4930
## Detection Rate : 0.3380
## Detection Prevalence : 0.5211
## Balanced Accuracy : 0.6623
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6.1_pred <- predict(smile__svm_model_6, tst_smile_boys)
summary(smile__svm_model_6.1_pred)
## spontaneous deliberate
## 33 44
smile__svm_model_6.1_confM <- confusionMatrix(
smile__svm_model_6.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 11
## deliberate 15 29
##
## Accuracy : 0.6623
## 95% CI : (0.5455, 0.7662)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.007868
##
## Kappa : 0.3209
##
## Mcnemar's Test P-Value : 0.556298
##
## Sensitivity : 0.5946
## Specificity : 0.7250
## Pos Pred Value : 0.6667
## Neg Pred Value : 0.6591
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.4286
## Balanced Accuracy : 0.6598
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6.2_pred <- predict(smile__svm_model_6, tst_smile_girls)
summary(smile__svm_model_6.2_pred)
## spontaneous deliberate
## 41 24
smile__svm_model_6.2_confM <- confusionMatrix(
smile__svm_model_6.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 15
## deliberate 7 17
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3203
##
## Mcnemar's Test P-Value : 0.135593
##
## Sensitivity : 0.7879
## Specificity : 0.5312
## Pos Pred Value : 0.6341
## Neg Pred Value : 0.7083
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.6308
## Balanced Accuracy : 0.6596
##
## 'Positive' Class : spontaneous
##
# model 6A dynamics and eye
set.seed(1973)
smile__svm_model_6A <-
train(smile_type ~ onset_mean + apex_mean + offset_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6A$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6509024 0.30311 0.1167325 0.2316699
summary(smile__svm_model_6A$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_6A_pred <- predict(smile__svm_model_6A, tst_smile)
summary(smile__svm_model_6A_pred)
## spontaneous deliberate
## 78 64
smile__svm_model_6A_confM <- confusionMatrix(
smile__svm_model_6A_pred,
tst_smile$smile_type
)
smile__svm_model_6A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 51 27
## deliberate 19 45
##
## Accuracy : 0.6761
## 95% CI : (0.5925, 0.7521)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 3.363e-05
##
## Kappa : 0.353
##
## Mcnemar's Test P-Value : 0.302
##
## Sensitivity : 0.7286
## Specificity : 0.6250
## Pos Pred Value : 0.6538
## Neg Pred Value : 0.7031
## Prevalence : 0.4930
## Detection Rate : 0.3592
## Detection Prevalence : 0.5493
## Balanced Accuracy : 0.6768
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6A.1_pred <- predict(smile__svm_model_6A, tst_smile_boys)
summary(smile__svm_model_6A.1_pred)
## spontaneous deliberate
## 37 40
smile__svm_model_6A.1_confM <- confusionMatrix(
smile__svm_model_6A.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 12
## deliberate 12 28
##
## Accuracy : 0.6883
## 95% CI : (0.5726, 0.7891)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.001955
##
## Kappa : 0.3757
##
## Mcnemar's Test P-Value : 1.000000
##
## Sensitivity : 0.6757
## Specificity : 0.7000
## Pos Pred Value : 0.6757
## Neg Pred Value : 0.7000
## Prevalence : 0.4805
## Detection Rate : 0.3247
## Detection Prevalence : 0.4805
## Balanced Accuracy : 0.6878
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6A.2_pred <- predict(smile__svm_model_6A, tst_smile_girls)
summary(smile__svm_model_6A.2_pred)
## spontaneous deliberate
## 41 24
smile__svm_model_6A.2_confM <- confusionMatrix(
smile__svm_model_6A.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 15
## deliberate 7 17
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3203
##
## Mcnemar's Test P-Value : 0.135593
##
## Sensitivity : 0.7879
## Specificity : 0.5312
## Pos Pred Value : 0.6341
## Neg Pred Value : 0.7083
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.6308
## Balanced Accuracy : 0.6596
##
## 'Positive' Class : spontaneous
##
# model 6B dynamics and lip
set.seed(1973)
smile__svm_model_6B <-
train(smile_type ~ onset_mean + apex_mean + offset_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6B$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6691789 0.3392816 0.07023856 0.1394647
summary(smile__svm_model_6B$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_6B_pred <- predict(smile__svm_model_6B, tst_smile)
summary(smile__svm_model_6B_pred)
## spontaneous deliberate
## 71 71
smile__svm_model_6B_confM <- confusionMatrix(
smile__svm_model_6B_pred,
tst_smile$smile_type
)
smile__svm_model_6B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 48 23
## deliberate 22 49
##
## Accuracy : 0.6831
## 95% CI : (0.5998, 0.7586)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 1.596e-05
##
## Kappa : 0.3662
##
## Mcnemar's Test P-Value : 1
##
## Sensitivity : 0.6857
## Specificity : 0.6806
## Pos Pred Value : 0.6761
## Neg Pred Value : 0.6901
## Prevalence : 0.4930
## Detection Rate : 0.3380
## Detection Prevalence : 0.5000
## Balanced Accuracy : 0.6831
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6B.1_pred <- predict(smile__svm_model_6B, tst_smile_boys)
summary(smile__svm_model_6B.1_pred)
## spontaneous deliberate
## 33 44
smile__svm_model_6B.1_confM <- confusionMatrix(
smile__svm_model_6B.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 10
## deliberate 14 30
##
## Accuracy : 0.6883
## 95% CI : (0.5726, 0.7891)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.001955
##
## Kappa : 0.3731
##
## Mcnemar's Test P-Value : 0.540291
##
## Sensitivity : 0.6216
## Specificity : 0.7500
## Pos Pred Value : 0.6970
## Neg Pred Value : 0.6818
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.4286
## Balanced Accuracy : 0.6858
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6B.2_pred <- predict(smile__svm_model_6B, tst_smile_girls)
summary(smile__svm_model_6B.2_pred)
## spontaneous deliberate
## 38 27
smile__svm_model_6B.2_confM <- confusionMatrix(
smile__svm_model_6B.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 13
## deliberate 8 19
##
## Accuracy : 0.6769
## 95% CI : (0.5495, 0.7877)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.004282
##
## Kappa : 0.3522
##
## Mcnemar's Test P-Value : 0.382733
##
## Sensitivity : 0.7576
## Specificity : 0.5938
## Pos Pred Value : 0.6579
## Neg Pred Value : 0.7037
## Prevalence : 0.5077
## Detection Rate : 0.3846
## Detection Prevalence : 0.5846
## Balanced Accuracy : 0.6757
##
## 'Positive' Class : spontaneous
##
# model 6C onset and movement
set.seed(1973)
smile__svm_model_6C <- train(smile_type ~ onset_mean + eye_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6C$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.660723 0.3212148 0.04400409 0.08849735
summary(smile__svm_model_6C$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_6C_pred <- predict(smile__svm_model_6C, tst_smile)
summary(smile__svm_model_6C_pred)
## spontaneous deliberate
## 77 65
smile__svm_model_6C_confM <- confusionMatrix(
smile__svm_model_6C_pred,
tst_smile$smile_type
)
smile__svm_model_6C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 42 35
## deliberate 28 37
##
## Accuracy : 0.5563
## 95% CI : (0.4707, 0.6396)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.1376
##
## Kappa : 0.1137
##
## Mcnemar's Test P-Value : 0.4497
##
## Sensitivity : 0.6000
## Specificity : 0.5139
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5692
## Prevalence : 0.4930
## Detection Rate : 0.2958
## Detection Prevalence : 0.5423
## Balanced Accuracy : 0.5569
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6C.1_pred <- predict(smile__svm_model_6C, tst_smile_boys)
summary(smile__svm_model_6C.1_pred)
## spontaneous deliberate
## 36 41
smile__svm_model_6C.1_confM <- confusionMatrix(
smile__svm_model_6C.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 15
## deliberate 16 25
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.1928
##
## Mcnemar's Test P-Value : 1.0000
##
## Sensitivity : 0.5676
## Specificity : 0.6250
## Pos Pred Value : 0.5833
## Neg Pred Value : 0.6098
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.4675
## Balanced Accuracy : 0.5963
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6C.2_pred <- predict(smile__svm_model_6C, tst_smile_girls)
summary(smile__svm_model_6C.2_pred)
## spontaneous deliberate
## 41 24
smile__svm_model_6C.2_confM <- confusionMatrix(
smile__svm_model_6C.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 20
## deliberate 12 12
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0114
##
## Mcnemar's Test P-Value : 0.2159
##
## Sensitivity : 0.6364
## Specificity : 0.3750
## Pos Pred Value : 0.5122
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3231
## Detection Prevalence : 0.6308
## Balanced Accuracy : 0.5057
##
## 'Positive' Class : spontaneous
##
# model 6D onset + eye
set.seed(1973)
smile__svm_model_6D <- train(smile_type ~ onset_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6D$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5433155 0.08600491 0.1022142 0.206062
summary(smile__svm_model_6D$finalModel)
## Length Class Mode
## 1 ksvm S4
kernlab::plot(smile__svm_model_6D$finalModel)
smile__svm_model_6D_pred <- predict(smile__svm_model_6D, tst_smile)
summary(smile__svm_model_6D_pred)
## spontaneous deliberate
## 60 82
smile__svm_model_6D_confM <- confusionMatrix(
smile__svm_model_6D_pred,
tst_smile$smile_type
)
smile__svm_model_6D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 33 27
## deliberate 37 45
##
## Accuracy : 0.5493
## 95% CI : (0.4636, 0.6328)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.1780
##
## Kappa : 0.0966
##
## Mcnemar's Test P-Value : 0.2606
##
## Sensitivity : 0.4714
## Specificity : 0.6250
## Pos Pred Value : 0.5500
## Neg Pred Value : 0.5488
## Prevalence : 0.4930
## Detection Rate : 0.2324
## Detection Prevalence : 0.4225
## Balanced Accuracy : 0.5482
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6D.1_pred <- predict(smile__svm_model_6D, tst_smile_boys)
summary(smile__svm_model_6D.1_pred)
## spontaneous deliberate
## 27 50
smile__svm_model_6D.1_confM <- confusionMatrix(
smile__svm_model_6D.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 14 13
## deliberate 23 27
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.0539
##
## Mcnemar's Test P-Value : 0.1336
##
## Sensitivity : 0.3784
## Specificity : 0.6750
## Pos Pred Value : 0.5185
## Neg Pred Value : 0.5400
## Prevalence : 0.4805
## Detection Rate : 0.1818
## Detection Prevalence : 0.3506
## Balanced Accuracy : 0.5267
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6D.2_pred <- predict(smile__svm_model_6D, tst_smile_girls)
summary(smile__svm_model_6D.2_pred)
## spontaneous deliberate
## 33 32
smile__svm_model_6D.2_confM <- confusionMatrix(
smile__svm_model_6D.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 14
## deliberate 14 18
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.1927
##
## Kappa : 0.1383
##
## Mcnemar's Test P-Value : 1.0000
##
## Sensitivity : 0.5758
## Specificity : 0.5625
## Pos Pred Value : 0.5758
## Neg Pred Value : 0.5625
## Prevalence : 0.5077
## Detection Rate : 0.2923
## Detection Prevalence : 0.5077
## Balanced Accuracy : 0.5691
##
## 'Positive' Class : spontaneous
##
# model 6E onset + lip
set.seed(1973)
smile__svm_model_6E <- train(smile_type ~ onset_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6E$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6483344 0.2962302 0.06798322 0.1368051
summary(smile__svm_model_6E$finalModel)
## Length Class Mode
## 1 ksvm S4
kernlab::plot(smile__svm_model_6E$finalModel)
smile__svm_model_6E_pred <- predict(smile__svm_model_6E, tst_smile)
summary(smile__svm_model_6E_pred)
## spontaneous deliberate
## 77 65
smile__svm_model_6E_confM <- confusionMatrix(
smile__svm_model_6E_pred,
tst_smile$smile_type
)
print(smile__svm_model_6E_confM)
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 43 34
## deliberate 27 38
##
## Accuracy : 0.5704
## 95% CI : (0.4847, 0.6531)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.07664
##
## Kappa : 0.1419
##
## Mcnemar's Test P-Value : 0.44236
##
## Sensitivity : 0.6143
## Specificity : 0.5278
## Pos Pred Value : 0.5584
## Neg Pred Value : 0.5846
## Prevalence : 0.4930
## Detection Rate : 0.3028
## Detection Prevalence : 0.5423
## Balanced Accuracy : 0.5710
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6E.1_pred <- predict(smile__svm_model_6E, tst_smile_boys)
summary(smile__svm_model_6E.1_pred)
## spontaneous deliberate
## 36 41
smile__svm_model_6E.1_confM <- confusionMatrix(
smile__svm_model_6E.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 15
## deliberate 16 25
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.1928
##
## Mcnemar's Test P-Value : 1.0000
##
## Sensitivity : 0.5676
## Specificity : 0.6250
## Pos Pred Value : 0.5833
## Neg Pred Value : 0.6098
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.4675
## Balanced Accuracy : 0.5963
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6E.2_pred <- predict(smile__svm_model_6E, tst_smile_girls)
summary(smile__svm_model_6E.2_pred)
## spontaneous deliberate
## 41 24
smile__svm_model_6E.2_confM <- confusionMatrix(
smile__svm_model_6E.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 19
## deliberate 11 13
##
## Accuracy : 0.5385
## 95% CI : (0.4103, 0.663)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.3553
##
## Kappa : 0.0732
##
## Mcnemar's Test P-Value : 0.2012
##
## Sensitivity : 0.6667
## Specificity : 0.4062
## Pos Pred Value : 0.5366
## Neg Pred Value : 0.5417
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.6308
## Balanced Accuracy : 0.5365
##
## 'Positive' Class : spontaneous
##
# model 6F apex + movements
set.seed(1973)
smile__svm_model_6F <- train(smile_type ~ apex_mean + eye_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6F$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5406473 0.08103459 0.09625901 0.19191
summary(smile__svm_model_6F$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_6F_pred <- predict(smile__svm_model_6F, tst_smile)
summary(smile__svm_model_6F_pred)
## spontaneous deliberate
## 70 72
smile__svm_model_6F_confM <- confusionMatrix(
smile__svm_model_6F_pred,
tst_smile$smile_type
)
smile__svm_model_6F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 41 29
## deliberate 29 43
##
## Accuracy : 0.5915
## 95% CI : (0.506, 0.6732)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.02653
##
## Kappa : 0.1829
##
## Mcnemar's Test P-Value : 1.00000
##
## Sensitivity : 0.5857
## Specificity : 0.5972
## Pos Pred Value : 0.5857
## Neg Pred Value : 0.5972
## Prevalence : 0.4930
## Detection Rate : 0.2887
## Detection Prevalence : 0.4930
## Balanced Accuracy : 0.5915
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6F.1_pred <- predict(smile__svm_model_6F, tst_smile_boys)
summary(smile__svm_model_6F.1_pred)
## spontaneous deliberate
## 38 39
smile__svm_model_6F.1_confM <- confusionMatrix(
smile__svm_model_6F.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 17
## deliberate 16 23
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.2126
##
## Kappa : 0.1424
##
## Mcnemar's Test P-Value : 1.0000
##
## Sensitivity : 0.5676
## Specificity : 0.5750
## Pos Pred Value : 0.5526
## Neg Pred Value : 0.5897
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.4935
## Balanced Accuracy : 0.5713
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6F.2_pred <- predict(smile__svm_model_6F, tst_smile_girls)
summary(smile__svm_model_6F.2_pred)
## spontaneous deliberate
## 32 33
smile__svm_model_6F.2_confM <- confusionMatrix(
smile__svm_model_6F.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 12
## deliberate 13 20
##
## Accuracy : 0.6154
## 95% CI : (0.4864, 0.7335)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.05294
##
## Kappa : 0.231
##
## Mcnemar's Test P-Value : 1.00000
##
## Sensitivity : 0.6061
## Specificity : 0.6250
## Pos Pred Value : 0.6250
## Neg Pred Value : 0.6061
## Prevalence : 0.5077
## Detection Rate : 0.3077
## Detection Prevalence : 0.4923
## Balanced Accuracy : 0.6155
##
## 'Positive' Class : spontaneous
##
# model 6G apex + eye
set.seed(1973)
smile__svm_model_6G <- train(smile_type ~ apex_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6G$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5381684 0.07448867 0.06085767 0.1208593
summary(smile__svm_model_6G$finalModel)
## Length Class Mode
## 1 ksvm S4
plot(smile__svm_model_6G$finalModel)
smile__svm_model_6G_pred <- predict(smile__svm_model_6G, tst_smile)
summary(smile__svm_model_6G_pred)
## spontaneous deliberate
## 51 91
smile__svm_model_6G_confM <- confusionMatrix(
smile__svm_model_6G_pred,
tst_smile$smile_type
)
smile__svm_model_6G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 25
## deliberate 44 47
##
## Accuracy : 0.5141
## 95% CI : (0.4288, 0.5987)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.46673
##
## Kappa : 0.0243
##
## Mcnemar's Test P-Value : 0.03024
##
## Sensitivity : 0.3714
## Specificity : 0.6528
## Pos Pred Value : 0.5098
## Neg Pred Value : 0.5165
## Prevalence : 0.4930
## Detection Rate : 0.1831
## Detection Prevalence : 0.3592
## Balanced Accuracy : 0.5121
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6G.1_pred <- predict(smile__svm_model_6G, tst_smile_boys)
summary(smile__svm_model_6G.1_pred)
## spontaneous deliberate
## 29 48
smile__svm_model_6G.1_confM <- confusionMatrix(
smile__svm_model_6G.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 13
## deliberate 21 27
##
## Accuracy : 0.5584
## 95% CI : (0.4407, 0.6716)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.2847
##
## Kappa : 0.1083
##
## Mcnemar's Test P-Value : 0.2299
##
## Sensitivity : 0.4324
## Specificity : 0.6750
## Pos Pred Value : 0.5517
## Neg Pred Value : 0.5625
## Prevalence : 0.4805
## Detection Rate : 0.2078
## Detection Prevalence : 0.3766
## Balanced Accuracy : 0.5537
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6G.2_pred <- predict(smile__svm_model_6G, tst_smile_girls)
summary(smile__svm_model_6G.2_pred)
## spontaneous deliberate
## 22 43
smile__svm_model_6G.2_confM <- confusionMatrix(
smile__svm_model_6G.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 10 12
## deliberate 23 20
##
## Accuracy : 0.4615
## 95% CI : (0.337, 0.5897)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.80736
##
## Kappa : -0.0716
##
## Mcnemar's Test P-Value : 0.09097
##
## Sensitivity : 0.3030
## Specificity : 0.6250
## Pos Pred Value : 0.4545
## Neg Pred Value : 0.4651
## Prevalence : 0.5077
## Detection Rate : 0.1538
## Detection Prevalence : 0.3385
## Balanced Accuracy : 0.4640
##
## 'Positive' Class : spontaneous
##
# model 6H apex + lip
set.seed(1973)
smile__svm_model_6H <- train(smile_type ~ apex_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6H$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5642714 0.1292836 0.0968237 0.1914572
summary(smile__svm_model_6H$finalModel)
## Length Class Mode
## 1 ksvm S4
kernlab::plot(smile__svm_model_6H$finalModel)
smile__svm_model_6H_pred <- predict(smile__svm_model_6H, tst_smile)
summary(smile__svm_model_6H_pred)
## spontaneous deliberate
## 79 63
smile__svm_model_6H_confM <- confusionMatrix(
smile__svm_model_6H_pred,
tst_smile$smile_type
)
smile__svm_model_6H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 33
## deliberate 24 39
##
## Accuracy : 0.5986
## 95% CI : (0.5131, 0.6799)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0177
##
## Kappa : 0.1985
##
## Mcnemar's Test P-Value : 0.2893
##
## Sensitivity : 0.6571
## Specificity : 0.5417
## Pos Pred Value : 0.5823
## Neg Pred Value : 0.6190
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.5563
## Balanced Accuracy : 0.5994
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6H.1_pred <- predict(smile__svm_model_6H, tst_smile_boys)
summary(smile__svm_model_6H.1_pred)
## spontaneous deliberate
## 40 37
smile__svm_model_6H.1_confM <- confusionMatrix(
smile__svm_model_6H.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 17
## deliberate 14 23
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.196
##
## Mcnemar's Test P-Value : 0.7194
##
## Sensitivity : 0.6216
## Specificity : 0.5750
## Pos Pred Value : 0.5750
## Neg Pred Value : 0.6216
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.5195
## Balanced Accuracy : 0.5983
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6H.2_pred <- predict(smile__svm_model_6H, tst_smile_girls)
summary(smile__svm_model_6H.2_pred)
## spontaneous deliberate
## 39 26
smile__svm_model_6H.2_confM <- confusionMatrix(
smile__svm_model_6H.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 16
## deliberate 10 16
##
## Accuracy : 0.6
## 95% CI : (0.471, 0.7196)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.08588
##
## Kappa : 0.1975
##
## Mcnemar's Test P-Value : 0.32680
##
## Sensitivity : 0.6970
## Specificity : 0.5000
## Pos Pred Value : 0.5897
## Neg Pred Value : 0.6154
## Prevalence : 0.5077
## Detection Rate : 0.3538
## Detection Prevalence : 0.6000
## Balanced Accuracy : 0.5985
##
## 'Positive' Class : spontaneous
##
# model 6I offset and movement
set.seed(1973)
smile__svm_model_6I <- train(smile_type ~ offset_mean + eye_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6I$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6662377 0.3343434 0.07243843 0.1434402
summary(smile__svm_model_6I$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_6I_pred <- predict(smile__svm_model_6I, tst_smile)
summary(smile__svm_model_6I_pred)
## spontaneous deliberate
## 82 60
smile__svm_model_6I_confM <- confusionMatrix(
smile__svm_model_6I_pred,
tst_smile$smile_type
)
smile__svm_model_6I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 51 31
## deliberate 19 41
##
## Accuracy : 0.6479
## 95% CI : (0.5634, 0.7261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0004898
##
## Kappa : 0.2973
##
## Mcnemar's Test P-Value : 0.1197949
##
## Sensitivity : 0.7286
## Specificity : 0.5694
## Pos Pred Value : 0.6220
## Neg Pred Value : 0.6833
## Prevalence : 0.4930
## Detection Rate : 0.3592
## Detection Prevalence : 0.5775
## Balanced Accuracy : 0.6490
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6I.1_pred <- predict(smile__svm_model_6I, tst_smile_boys)
summary(smile__svm_model_6I.1_pred)
## spontaneous deliberate
## 38 39
smile__svm_model_6I.1_confM <- confusionMatrix(
smile__svm_model_6I.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 13
## deliberate 12 27
##
## Accuracy : 0.6753
## 95% CI : (0.559, 0.7777)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.00403
##
## Kappa : 0.3503
##
## Mcnemar's Test P-Value : 1.00000
##
## Sensitivity : 0.6757
## Specificity : 0.6750
## Pos Pred Value : 0.6579
## Neg Pred Value : 0.6923
## Prevalence : 0.4805
## Detection Rate : 0.3247
## Detection Prevalence : 0.4935
## Balanced Accuracy : 0.6753
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6I.2_pred <- predict(smile__svm_model_6I, tst_smile_girls)
summary(smile__svm_model_6I.2_pred)
## spontaneous deliberate
## 44 21
smile__svm_model_6I.2_confM <- confusionMatrix(
smile__svm_model_6I.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 18
## deliberate 7 14
##
## Accuracy : 0.6154
## 95% CI : (0.4864, 0.7335)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.05294
##
## Kappa : 0.2266
##
## Mcnemar's Test P-Value : 0.04550
##
## Sensitivity : 0.7879
## Specificity : 0.4375
## Pos Pred Value : 0.5909
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.6769
## Balanced Accuracy : 0.6127
##
## 'Positive' Class : spontaneous
##
# model 6J offset + eye
set.seed(1973)
smile__svm_model_6J <- train(smile_type ~ offset_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6J$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.552217 0.1031459 0.1055909 0.2110006
summary(smile__svm_model_6J$finalModel)
## Length Class Mode
## 1 ksvm S4
kernlab::plot(smile__svm_model_6J$finalModel)
smile__svm_model_6J_pred <- predict(smile__svm_model_6J, tst_smile)
summary(smile__svm_model_6J_pred)
## spontaneous deliberate
## 62 80
smile__svm_model_6J_confM <- confusionMatrix(
smile__svm_model_6J_pred,
tst_smile$smile_type
)
smile__svm_model_6J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 40 22
## deliberate 30 50
##
## Accuracy : 0.6338
## 95% CI : (0.5489, 0.713)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.001568
##
## Kappa : 0.2663
##
## Mcnemar's Test P-Value : 0.331685
##
## Sensitivity : 0.5714
## Specificity : 0.6944
## Pos Pred Value : 0.6452
## Neg Pred Value : 0.6250
## Prevalence : 0.4930
## Detection Rate : 0.2817
## Detection Prevalence : 0.4366
## Balanced Accuracy : 0.6329
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6J.1_pred <- predict(smile__svm_model_6J, tst_smile_boys)
summary(smile__svm_model_6J.1_pred)
## spontaneous deliberate
## 30 47
smile__svm_model_6J.1_confM <- confusionMatrix(
smile__svm_model_6J.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 18 12
## deliberate 19 28
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.1878
##
## Mcnemar's Test P-Value : 0.2812
##
## Sensitivity : 0.4865
## Specificity : 0.7000
## Pos Pred Value : 0.6000
## Neg Pred Value : 0.5957
## Prevalence : 0.4805
## Detection Rate : 0.2338
## Detection Prevalence : 0.3896
## Balanced Accuracy : 0.5932
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6J.2_pred <- predict(smile__svm_model_6J, tst_smile_girls)
summary(smile__svm_model_6J.2_pred)
## spontaneous deliberate
## 32 33
smile__svm_model_6J.2_confM <- confusionMatrix(
smile__svm_model_6J.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 10
## deliberate 11 22
##
## Accuracy : 0.6769
## 95% CI : (0.5495, 0.7877)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.004282
##
## Kappa : 0.354
##
## Mcnemar's Test P-Value : 1.000000
##
## Sensitivity : 0.6667
## Specificity : 0.6875
## Pos Pred Value : 0.6875
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.4923
## Balanced Accuracy : 0.6771
##
## 'Positive' Class : spontaneous
##
# model 6K offset + lip
set.seed(1973)
smile__svm_model_6K <- train(smile_type ~ offset_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_6K
## Support Vector Machines with Linear Kernel
##
## 333 samples
## 2 predictor
## 2 classes: 'spontaneous', 'deliberate '
##
## No pre-processing
## Resampling: Cross-Validated (10 fold)
## Summary of sample sizes: 300, 300, 300, 299, 299, 300, ...
## Resampling results:
##
## Accuracy Kappa
## 0.6787322 0.3589939
##
## Tuning parameter 'C' was held constant at a value of 1
smile__svm_model_6K$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6787322 0.3589939 0.07408875 0.1473768
summary(smile__svm_model_6K$finalModel)
## Length Class Mode
## 1 ksvm S4
kernlab::plot(smile__svm_model_6K$finalModel)
smile__svm_model_6K$bestTune
## C
## 1 1
smile__svm_model_6K_pred <- predict(smile__svm_model_6K, tst_smile)
summary(smile__svm_model_6K_pred)
## spontaneous deliberate
## 83 59
smile__svm_model_6K_confM <- confusionMatrix(
smile__svm_model_6K_pred,
tst_smile$smile_type
)
smile__svm_model_6K_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 53 30
## deliberate 17 42
##
## Accuracy : 0.669
## 95% CI : (0.5852, 0.7456)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 6.871e-05
##
## Kappa : 0.3396
##
## Mcnemar's Test P-Value : 0.08005
##
## Sensitivity : 0.7571
## Specificity : 0.5833
## Pos Pred Value : 0.6386
## Neg Pred Value : 0.7119
## Prevalence : 0.4930
## Detection Rate : 0.3732
## Detection Prevalence : 0.5845
## Balanced Accuracy : 0.6702
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_6K.1_pred <- predict(smile__svm_model_6K, tst_smile_boys)
summary(smile__svm_model_6K.1_pred)
## spontaneous deliberate
## 42 35
smile__svm_model_6K.1_confM <- confusionMatrix(
smile__svm_model_6K.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_6K.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 27 15
## deliberate 10 25
##
## Accuracy : 0.6753
## 95% CI : (0.559, 0.7777)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.00403
##
## Kappa : 0.3529
##
## Mcnemar's Test P-Value : 0.42371
##
## Sensitivity : 0.7297
## Specificity : 0.6250
## Pos Pred Value : 0.6429
## Neg Pred Value : 0.7143
## Prevalence : 0.4805
## Detection Rate : 0.3506
## Detection Prevalence : 0.5455
## Balanced Accuracy : 0.6774
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_6K.2_pred <- predict(smile__svm_model_6K, tst_smile_girls)
summary(smile__svm_model_6K.2_pred)
## spontaneous deliberate
## 41 24
smile__svm_model_6K.2_confM <- confusionMatrix(
smile__svm_model_6K.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_6K.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 15
## deliberate 7 17
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3203
##
## Mcnemar's Test P-Value : 0.135593
##
## Sensitivity : 0.7879
## Specificity : 0.5312
## Pos Pred Value : 0.6341
## Neg Pred Value : 0.7083
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.6308
## Balanced Accuracy : 0.6596
##
## 'Positive' Class : spontaneous
##
# dynamics and AU's
set.seed(1973)
smile__svm_model_7 <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.7498552 0.4990742 0.1107154 0.221964
summary(smile__svm_model_7$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7_pred <- predict(smile__svm_model_7, tst_smile)
summary(smile__svm_model_7_pred)
## spontaneous deliberate
## 68 74
smile__svm_model_7_confM <- confusionMatrix(
smile__svm_model_7_pred,
tst_smile$smile_type
)
smile__svm_model_7_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 46 22
## deliberate 24 50
##
## Accuracy : 0.6761
## 95% CI : (0.5925, 0.7521)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 3.363e-05
##
## Kappa : 0.3517
##
## Mcnemar's Test P-Value : 0.8828
##
## Sensitivity : 0.6571
## Specificity : 0.6944
## Pos Pred Value : 0.6765
## Neg Pred Value : 0.6757
## Prevalence : 0.4930
## Detection Rate : 0.3239
## Detection Prevalence : 0.4789
## Balanced Accuracy : 0.6758
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7.1_pred <- predict(smile__svm_model_7, tst_smile_boys)
summary(smile__svm_model_7.1_pred)
## spontaneous deliberate
## 35 42
smile__svm_model_7.1_confM <- confusionMatrix(
smile__svm_model_7.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 11
## deliberate 13 29
##
## Accuracy : 0.6883
## 95% CI : (0.5726, 0.7891)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.001955
##
## Kappa : 0.3744
##
## Mcnemar's Test P-Value : 0.838256
##
## Sensitivity : 0.6486
## Specificity : 0.7250
## Pos Pred Value : 0.6857
## Neg Pred Value : 0.6905
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.4545
## Balanced Accuracy : 0.6868
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7.2_pred <- predict(smile__svm_model_7, tst_smile_girls)
summary(smile__svm_model_7.2_pred)
## spontaneous deliberate
## 33 32
smile__svm_model_7.2_confM <- confusionMatrix(
smile__svm_model_7.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 11
## deliberate 11 21
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3229
##
## Mcnemar's Test P-Value : 1.000000
##
## Sensitivity : 0.6667
## Specificity : 0.6562
## Pos Pred Value : 0.6667
## Neg Pred Value : 0.6562
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.5077
## Balanced Accuracy : 0.6615
##
## 'Positive' Class : spontaneous
##
# 7A AU's and onset
set.seed(1973)
smile__svm_model_7A <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
onset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7A$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6871992 0.3758363 0.1064631 0.210235
summary(smile__svm_model_7A$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7A_pred <- predict(smile__svm_model_7A, tst_smile)
summary(smile__svm_model_7A_pred)
## spontaneous deliberate
## 61 81
smile__svm_model_7A_confM <- confusionMatrix(
smile__svm_model_7A_pred,
tst_smile$smile_type
)
smile__svm_model_7A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 36 25
## deliberate 34 47
##
## Accuracy : 0.5845
## 95% CI : (0.4989, 0.6665)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.03874
##
## Kappa : 0.1674
##
## Mcnemar's Test P-Value : 0.29764
##
## Sensitivity : 0.5143
## Specificity : 0.6528
## Pos Pred Value : 0.5902
## Neg Pred Value : 0.5802
## Prevalence : 0.4930
## Detection Rate : 0.2535
## Detection Prevalence : 0.4296
## Balanced Accuracy : 0.5835
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7A.1_pred <- predict(smile__svm_model_7A, tst_smile_boys)
summary(smile__svm_model_7A.1_pred)
## spontaneous deliberate
## 33 44
smile__svm_model_7A.1_confM <- confusionMatrix(
smile__svm_model_7A.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 13
## deliberate 17 27
##
## Accuracy : 0.6104
## 95% CI : (0.4925, 0.7195)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.06859
##
## Kappa : 0.2164
##
## Mcnemar's Test P-Value : 0.58388
##
## Sensitivity : 0.5405
## Specificity : 0.6750
## Pos Pred Value : 0.6061
## Neg Pred Value : 0.6136
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.4286
## Balanced Accuracy : 0.6078
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7A.2_pred <- predict(smile__svm_model_7A, tst_smile_girls)
summary(smile__svm_model_7A.2_pred)
## spontaneous deliberate
## 28 37
smile__svm_model_7A.2_confM <- confusionMatrix(
smile__svm_model_7A.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 12
## deliberate 17 20
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.2678
##
## Kappa : 0.1096
##
## Mcnemar's Test P-Value : 0.4576
##
## Sensitivity : 0.4848
## Specificity : 0.6250
## Pos Pred Value : 0.5714
## Neg Pred Value : 0.5405
## Prevalence : 0.5077
## Detection Rate : 0.2462
## Detection Prevalence : 0.4308
## Balanced Accuracy : 0.5549
##
## 'Positive' Class : spontaneous
##
# 7B AU's and apex
set.seed(1973)
smile__svm_model_7B <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
apex_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7B$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6873886 0.3751691 0.1010571 0.2013494
summary(smile__svm_model_7B$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7B_pred <- predict(smile__svm_model_7B, tst_smile)
summary(smile__svm_model_7B_pred)
## spontaneous deliberate
## 60 82
smile__svm_model_7B_confM <- confusionMatrix(
smile__svm_model_7B_pred,
tst_smile$smile_type
)
smile__svm_model_7B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 35 25
## deliberate 35 47
##
## Accuracy : 0.5775
## 95% CI : (0.4918, 0.6598)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.05517
##
## Kappa : 0.1531
##
## Mcnemar's Test P-Value : 0.24528
##
## Sensitivity : 0.5000
## Specificity : 0.6528
## Pos Pred Value : 0.5833
## Neg Pred Value : 0.5732
## Prevalence : 0.4930
## Detection Rate : 0.2465
## Detection Prevalence : 0.4225
## Balanced Accuracy : 0.5764
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7B.1_pred <- predict(smile__svm_model_7B, tst_smile_boys)
summary(smile__svm_model_7B.1_pred)
## spontaneous deliberate
## 32 45
smile__svm_model_7B.1_confM <- confusionMatrix(
smile__svm_model_7B.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 13
## deliberate 18 27
##
## Accuracy : 0.5974
## 95% CI : (0.4794, 0.7077)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1045
##
## Kappa : 0.1895
##
## Mcnemar's Test P-Value : 0.4725
##
## Sensitivity : 0.5135
## Specificity : 0.6750
## Pos Pred Value : 0.5938
## Neg Pred Value : 0.6000
## Prevalence : 0.4805
## Detection Rate : 0.2468
## Detection Prevalence : 0.4156
## Balanced Accuracy : 0.5943
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7B.2_pred <- predict(smile__svm_model_7B, tst_smile_girls)
summary(smile__svm_model_7B.2_pred)
## spontaneous deliberate
## 28 37
smile__svm_model_7B.2_confM <- confusionMatrix(
smile__svm_model_7B.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 12
## deliberate 17 20
##
## Accuracy : 0.5538
## 95% CI : (0.4253, 0.6773)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.2678
##
## Kappa : 0.1096
##
## Mcnemar's Test P-Value : 0.4576
##
## Sensitivity : 0.4848
## Specificity : 0.6250
## Pos Pred Value : 0.5714
## Neg Pred Value : 0.5405
## Prevalence : 0.5077
## Detection Rate : 0.2462
## Detection Prevalence : 0.4308
## Balanced Accuracy : 0.5549
##
## 'Positive' Class : spontaneous
##
# 7C AU's and offset
set.seed(1973)
smile__svm_model_7C <- train(smile_type ~ AU01_r_mean + AU02_r_mean +
AU04_r_mean + AU05_r_mean + AU06_r_mean +
AU07_r_mean + AU09_r_mean + AU10_r_mean +
AU12_r_mean + AU14_r_mean + AU15_r_mean +
AU17_r_mean + AU20_r_mean + AU23_r_mean +
AU25_r_mean + AU26_r_mean + AU45_r_mean +
offset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7C$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6780136 0.3569963 0.09684482 0.1925157
summary(smile__svm_model_7C$finalModel)
## Length Class Mode
## 1 ksvm S4
# visualize the svm using the rattle package
# plot(smile__svm_model_7C$finalModel)
smile__svm_model_7C_pred <- predict(smile__svm_model_7C, tst_smile)
summary(smile__svm_model_7C_pred)
## spontaneous deliberate
## 64 78
smile__svm_model_7C_confM <- confusionMatrix(
smile__svm_model_7C_pred,
tst_smile$smile_type
)
smile__svm_model_7C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 39 25
## deliberate 31 47
##
## Accuracy : 0.6056
## 95% CI : (0.5202, 0.6865)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.01151
##
## Kappa : 0.2102
##
## Mcnemar's Test P-Value : 0.50404
##
## Sensitivity : 0.5571
## Specificity : 0.6528
## Pos Pred Value : 0.6094
## Neg Pred Value : 0.6026
## Prevalence : 0.4930
## Detection Rate : 0.2746
## Detection Prevalence : 0.4507
## Balanced Accuracy : 0.6050
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7C.1_pred <- predict(smile__svm_model_7C, tst_smile_boys)
summary(smile__svm_model_7C.1_pred)
## spontaneous deliberate
## 34 43
smile__svm_model_7C.1_confM <- confusionMatrix(
smile__svm_model_7C.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 13
## deliberate 16 27
##
## Accuracy : 0.6234
## 95% CI : (0.5056, 0.7313)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.04297
##
## Kappa : 0.2433
##
## Mcnemar's Test P-Value : 0.71035
##
## Sensitivity : 0.5676
## Specificity : 0.6750
## Pos Pred Value : 0.6176
## Neg Pred Value : 0.6279
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.4416
## Balanced Accuracy : 0.6213
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7C.2_pred <- predict(smile__svm_model_7C, tst_smile_girls)
summary(smile__svm_model_7C.2_pred)
## spontaneous deliberate
## 30 35
smile__svm_model_7C.2_confM <- confusionMatrix(
smile__svm_model_7C.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 18 12
## deliberate 15 20
##
## Accuracy : 0.5846
## 95% CI : (0.4556, 0.7056)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.1320
##
## Kappa : 0.1702
##
## Mcnemar's Test P-Value : 0.7003
##
## Sensitivity : 0.5455
## Specificity : 0.6250
## Pos Pred Value : 0.6000
## Neg Pred Value : 0.5714
## Prevalence : 0.5077
## Detection Rate : 0.2769
## Detection Prevalence : 0.4615
## Balanced Accuracy : 0.5852
##
## 'Positive' Class : spontaneous
##
# 7D AU's selection and dynamics
set.seed(1973)
smile__svm_model_7D <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7D$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6690842 0.3379967 0.08313076 0.1670725
summary(smile__svm_model_7D$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7D_pred <- predict(smile__svm_model_7D, tst_smile)
summary(smile__svm_model_7D_pred)
## spontaneous deliberate
## 60 82
smile__svm_model_7D_confM <- confusionMatrix(
smile__svm_model_7D_pred,
tst_smile$smile_type
)
smile__svm_model_7D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 42 18
## deliberate 28 54
##
## Accuracy : 0.6761
## 95% CI : (0.5925, 0.7521)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 3.363e-05
##
## Kappa : 0.3507
##
## Mcnemar's Test P-Value : 0.1845
##
## Sensitivity : 0.6000
## Specificity : 0.7500
## Pos Pred Value : 0.7000
## Neg Pred Value : 0.6585
## Prevalence : 0.4930
## Detection Rate : 0.2958
## Detection Prevalence : 0.4225
## Balanced Accuracy : 0.6750
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7D.1_pred <- predict(smile__svm_model_7D, tst_smile_boys)
summary(smile__svm_model_7D.1_pred)
## spontaneous deliberate
## 25 52
smile__svm_model_7D.1_confM <- confusionMatrix(
smile__svm_model_7D.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 6
## deliberate 18 34
##
## Accuracy : 0.6883
## 95% CI : (0.5726, 0.7891)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.001955
##
## Kappa : 0.368
##
## Mcnemar's Test P-Value : 0.024745
##
## Sensitivity : 0.5135
## Specificity : 0.8500
## Pos Pred Value : 0.7600
## Neg Pred Value : 0.6538
## Prevalence : 0.4805
## Detection Rate : 0.2468
## Detection Prevalence : 0.3247
## Balanced Accuracy : 0.6818
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7D.2_pred <- predict(smile__svm_model_7D, tst_smile_girls)
summary(smile__svm_model_7D.2_pred)
## spontaneous deliberate
## 35 30
smile__svm_model_7D.2_confM <- confusionMatrix(
smile__svm_model_7D.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 12
## deliberate 10 20
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3223
##
## Mcnemar's Test P-Value : 0.831170
##
## Sensitivity : 0.6970
## Specificity : 0.6250
## Pos Pred Value : 0.6571
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.3538
## Detection Prevalence : 0.5385
## Balanced Accuracy : 0.6610
##
## 'Positive' Class : spontaneous
##
# 7E AU's selection and onset
set.seed(1973)
smile__svm_model_7E <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7E$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.582002 0.1671039 0.06536007 0.131873
summary(smile__svm_model_7E$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7E_pred <- predict(smile__svm_model_7E, tst_smile)
summary(smile__svm_model_7E_pred)
## spontaneous deliberate
## 99 43
smile__svm_model_7E_confM <- confusionMatrix(
smile__svm_model_7E_pred,
tst_smile$smile_type
)
smile__svm_model_7E_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 51 48
## deliberate 19 24
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3375712
##
## Kappa : 0.0616
##
## Mcnemar's Test P-Value : 0.0006245
##
## Sensitivity : 0.7286
## Specificity : 0.3333
## Pos Pred Value : 0.5152
## Neg Pred Value : 0.5581
## Prevalence : 0.4930
## Detection Rate : 0.3592
## Detection Prevalence : 0.6972
## Balanced Accuracy : 0.5310
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7E.1_pred <- predict(smile__svm_model_7E, tst_smile_boys)
summary(smile__svm_model_7E.1_pred)
## spontaneous deliberate
## 51 26
smile__svm_model_7E.1_confM <- confusionMatrix(
smile__svm_model_7E.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 25
## deliberate 11 15
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.45523
##
## Kappa : 0.0766
##
## Mcnemar's Test P-Value : 0.03026
##
## Sensitivity : 0.7027
## Specificity : 0.3750
## Pos Pred Value : 0.5098
## Neg Pred Value : 0.5769
## Prevalence : 0.4805
## Detection Rate : 0.3377
## Detection Prevalence : 0.6623
## Balanced Accuracy : 0.5389
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7E.2_pred <- predict(smile__svm_model_7E, tst_smile_girls)
summary(smile__svm_model_7E.2_pred)
## spontaneous deliberate
## 48 17
smile__svm_model_7E.2_confM <- confusionMatrix(
smile__svm_model_7E.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 23
## deliberate 8 9
##
## Accuracy : 0.5231
## 95% CI : (0.3954, 0.6485)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.45095
##
## Kappa : 0.0391
##
## Mcnemar's Test P-Value : 0.01192
##
## Sensitivity : 0.7576
## Specificity : 0.2812
## Pos Pred Value : 0.5208
## Neg Pred Value : 0.5294
## Prevalence : 0.5077
## Detection Rate : 0.3846
## Detection Prevalence : 0.7385
## Balanced Accuracy : 0.5194
##
## 'Positive' Class : spontaneous
##
# 7B apex AU's selection and apex
set.seed(1973)
smile__svm_model_7F <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + apex_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7F$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5768549 0.1536795 0.07359117 0.1488513
summary(smile__svm_model_7F$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7F_pred <- predict(smile__svm_model_7F, tst_smile)
summary(smile__svm_model_7F_pred)
## spontaneous deliberate
## 73 69
smile__svm_model_7F_confM <- confusionMatrix(
smile__svm_model_7F_pred,
tst_smile$smile_type
)
smile__svm_model_7F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 38 35
## deliberate 32 37
##
## Accuracy : 0.5282
## 95% CI : (0.4427, 0.6124)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.3376
##
## Kappa : 0.0567
##
## Mcnemar's Test P-Value : 0.8070
##
## Sensitivity : 0.5429
## Specificity : 0.5139
## Pos Pred Value : 0.5205
## Neg Pred Value : 0.5362
## Prevalence : 0.4930
## Detection Rate : 0.2676
## Detection Prevalence : 0.5141
## Balanced Accuracy : 0.5284
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7F.1_pred <- predict(smile__svm_model_7F, tst_smile_boys)
summary(smile__svm_model_7F.1_pred)
## spontaneous deliberate
## 33 44
smile__svm_model_7F.1_confM <- confusionMatrix(
smile__svm_model_7F.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 14
## deliberate 18 26
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.1642
##
## Mcnemar's Test P-Value : 0.5959
##
## Sensitivity : 0.5135
## Specificity : 0.6500
## Pos Pred Value : 0.5758
## Neg Pred Value : 0.5909
## Prevalence : 0.4805
## Detection Rate : 0.2468
## Detection Prevalence : 0.4286
## Balanced Accuracy : 0.5818
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7F.2_pred <- predict(smile__svm_model_7F, tst_smile_girls)
summary(smile__svm_model_7F.2_pred)
## spontaneous deliberate
## 40 25
smile__svm_model_7F.2_confM <- confusionMatrix(
smile__svm_model_7F.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 21
## deliberate 14 11
##
## Accuracy : 0.4615
## 95% CI : (0.337, 0.5897)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.8074
##
## Kappa : -0.0808
##
## Mcnemar's Test P-Value : 0.3105
##
## Sensitivity : 0.5758
## Specificity : 0.3438
## Pos Pred Value : 0.4750
## Neg Pred Value : 0.4400
## Prevalence : 0.5077
## Detection Rate : 0.2923
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.4598
##
## 'Positive' Class : spontaneous
##
# 7G AU's selection and offset
set.seed(1973)
smile__svm_model_7G <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + offset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7G$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5761141 0.1547211 0.06019134 0.1219688
summary(smile__svm_model_7G$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7G_pred <- predict(smile__svm_model_7G, tst_smile)
summary(smile__svm_model_7G_pred)
## spontaneous deliberate
## 92 50
smile__svm_model_7G_confM <- confusionMatrix(
smile__svm_model_7G_pred,
tst_smile$smile_type
)
smile__svm_model_7G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 50 42
## deliberate 20 30
##
## Accuracy : 0.5634
## 95% CI : (0.4777, 0.6464)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.103915
##
## Kappa : 0.1304
##
## Mcnemar's Test P-Value : 0.007653
##
## Sensitivity : 0.7143
## Specificity : 0.4167
## Pos Pred Value : 0.5435
## Neg Pred Value : 0.6000
## Prevalence : 0.4930
## Detection Rate : 0.3521
## Detection Prevalence : 0.6479
## Balanced Accuracy : 0.5655
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7G.1_pred <- predict(smile__svm_model_7G, tst_smile_boys)
summary(smile__svm_model_7G.1_pred)
## spontaneous deliberate
## 49 28
smile__svm_model_7G.1_confM <- confusionMatrix(
smile__svm_model_7G.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 24
## deliberate 12 16
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.45523
##
## Kappa : 0.0748
##
## Mcnemar's Test P-Value : 0.06675
##
## Sensitivity : 0.6757
## Specificity : 0.4000
## Pos Pred Value : 0.5102
## Neg Pred Value : 0.5714
## Prevalence : 0.4805
## Detection Rate : 0.3247
## Detection Prevalence : 0.6364
## Balanced Accuracy : 0.5378
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7G.2_pred <- predict(smile__svm_model_7G, tst_smile_girls)
summary(smile__svm_model_7G.2_pred)
## spontaneous deliberate
## 43 22
smile__svm_model_7G.2_confM <- confusionMatrix(
smile__svm_model_7G.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 18
## deliberate 8 14
##
## Accuracy : 0.6
## 95% CI : (0.471, 0.7196)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.08588
##
## Kappa : 0.196
##
## Mcnemar's Test P-Value : 0.07756
##
## Sensitivity : 0.7576
## Specificity : 0.4375
## Pos Pred Value : 0.5814
## Neg Pred Value : 0.6364
## Prevalence : 0.5077
## Detection Rate : 0.3846
## Detection Prevalence : 0.6615
## Balanced Accuracy : 0.5975
##
## 'Positive' Class : spontaneous
##
# 7H selection AU's + temporal features
set.seed(1973)
smile__svm_model_7H <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7H$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6985851 0.3968121 0.09919617 0.1981549
summary(smile__svm_model_7H$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7H_pred <- predict(smile__svm_model_7H, tst_smile)
summary(smile__svm_model_7H_pred)
## spontaneous deliberate
## 59 83
smile__svm_model_7H_confM <- confusionMatrix(
smile__svm_model_7H_pred,
tst_smile$smile_type
)
smile__svm_model_7H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 43 16
## deliberate 27 56
##
## Accuracy : 0.6972
## 95% CI : (0.6145, 0.7714)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 3.274e-06
##
## Kappa : 0.3929
##
## Mcnemar's Test P-Value : 0.1273
##
## Sensitivity : 0.6143
## Specificity : 0.7778
## Pos Pred Value : 0.7288
## Neg Pred Value : 0.6747
## Prevalence : 0.4930
## Detection Rate : 0.3028
## Detection Prevalence : 0.4155
## Balanced Accuracy : 0.6960
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7H.1_pred <- predict(smile__svm_model_7H, tst_smile_boys)
summary(smile__svm_model_7H.1_pred)
## spontaneous deliberate
## 25 52
smile__svm_model_7H.1_confM <- confusionMatrix(
smile__svm_model_7H.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 5
## deliberate 17 35
##
## Accuracy : 0.7143
## 95% CI : (0.6, 0.8115)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.0003882
##
## Kappa : 0.4207
##
## Mcnemar's Test P-Value : 0.0190165
##
## Sensitivity : 0.5405
## Specificity : 0.8750
## Pos Pred Value : 0.8000
## Neg Pred Value : 0.6731
## Prevalence : 0.4805
## Detection Rate : 0.2597
## Detection Prevalence : 0.3247
## Balanced Accuracy : 0.7078
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7H.2_pred <- predict(smile__svm_model_7H, tst_smile_girls)
summary(smile__svm_model_7H.2_pred)
## spontaneous deliberate
## 34 31
smile__svm_model_7H.2_confM <- confusionMatrix(
smile__svm_model_7H.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 11
## deliberate 10 21
##
## Accuracy : 0.6769
## 95% CI : (0.5495, 0.7877)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.004282
##
## Kappa : 0.3534
##
## Mcnemar's Test P-Value : 1.000000
##
## Sensitivity : 0.6970
## Specificity : 0.6562
## Pos Pred Value : 0.6765
## Neg Pred Value : 0.6774
## Prevalence : 0.5077
## Detection Rate : 0.3538
## Detection Prevalence : 0.5231
## Balanced Accuracy : 0.6766
##
## 'Positive' Class : spontaneous
##
# 7I
set.seed(1973)
smile__svm_model_7I <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7I$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6392324 0.2786654 0.06036414 0.1196935
summary(smile__svm_model_7I$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7I_pred <- predict(smile__svm_model_7I, tst_smile)
summary(smile__svm_model_7I_pred)
## spontaneous deliberate
## 62 80
smile__svm_model_7I_confM <- confusionMatrix(
smile__svm_model_7I_pred,
tst_smile$smile_type
)
smile__svm_model_7I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 41 21
## deliberate 29 51
##
## Accuracy : 0.6479
## 95% CI : (0.5634, 0.7261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0004898
##
## Kappa : 0.2945
##
## Mcnemar's Test P-Value : 0.3221988
##
## Sensitivity : 0.5857
## Specificity : 0.7083
## Pos Pred Value : 0.6613
## Neg Pred Value : 0.6375
## Prevalence : 0.4930
## Detection Rate : 0.2887
## Detection Prevalence : 0.4366
## Balanced Accuracy : 0.6470
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7I.1_pred <- predict(smile__svm_model_7I, tst_smile_boys)
summary(smile__svm_model_7I.1_pred)
## spontaneous deliberate
## 27 50
smile__svm_model_7I.1_confM <- confusionMatrix(
smile__svm_model_7I.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 6
## deliberate 16 34
##
## Accuracy : 0.7143
## 95% CI : (0.6, 0.8115)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.0003882
##
## Kappa : 0.4218
##
## Mcnemar's Test P-Value : 0.0550088
##
## Sensitivity : 0.5676
## Specificity : 0.8500
## Pos Pred Value : 0.7778
## Neg Pred Value : 0.6800
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.3506
## Balanced Accuracy : 0.7088
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7I.2_pred <- predict(smile__svm_model_7I, tst_smile_girls)
summary(smile__svm_model_7I.2_pred)
## spontaneous deliberate
## 35 30
smile__svm_model_7I.2_confM <- confusionMatrix(
smile__svm_model_7I.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 15
## deliberate 13 17
##
## Accuracy : 0.5692
## 95% CI : (0.4404, 0.6915)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.1927
##
## Kappa : 0.1374
##
## Mcnemar's Test P-Value : 0.8501
##
## Sensitivity : 0.6061
## Specificity : 0.5312
## Pos Pred Value : 0.5714
## Neg Pred Value : 0.5667
## Prevalence : 0.5077
## Detection Rate : 0.3077
## Detection Prevalence : 0.5385
## Balanced Accuracy : 0.5687
##
## 'Positive' Class : spontaneous
##
# 7J
set.seed(1973)
smile__svm_model_7J <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
apex_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7J$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6424465 0.284497 0.07854931 0.1560307
summary(smile__svm_model_7J$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7J_pred <- predict(smile__svm_model_7J, tst_smile)
summary(smile__svm_model_7J_pred)
## spontaneous deliberate
## 65 77
smile__svm_model_7J_confM <- confusionMatrix(
smile__svm_model_7J_pred,
tst_smile$smile_type
)
smile__svm_model_7J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 43 22
## deliberate 27 50
##
## Accuracy : 0.6549
## 95% CI : (0.5706, 0.7326)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0002621
##
## Kappa : 0.309
##
## Mcnemar's Test P-Value : 0.5677092
##
## Sensitivity : 0.6143
## Specificity : 0.6944
## Pos Pred Value : 0.6615
## Neg Pred Value : 0.6494
## Prevalence : 0.4930
## Detection Rate : 0.3028
## Detection Prevalence : 0.4577
## Balanced Accuracy : 0.6544
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7J.1_pred <- predict(smile__svm_model_7J, tst_smile_boys)
summary(smile__svm_model_7J.1_pred)
## spontaneous deliberate
## 30 47
smile__svm_model_7J.1_confM <- confusionMatrix(
smile__svm_model_7J.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 6
## deliberate 13 34
##
## Accuracy : 0.7532
## 95% CI : (0.6418, 0.8444)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 2.185e-05
##
## Kappa : 0.5022
##
## Mcnemar's Test P-Value : 0.1687
##
## Sensitivity : 0.6486
## Specificity : 0.8500
## Pos Pred Value : 0.8000
## Neg Pred Value : 0.7234
## Prevalence : 0.4805
## Detection Rate : 0.3117
## Detection Prevalence : 0.3896
## Balanced Accuracy : 0.7493
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7J.2_pred <- predict(smile__svm_model_7J, tst_smile_girls)
summary(smile__svm_model_7J.2_pred)
## spontaneous deliberate
## 35 30
smile__svm_model_7J.2_confM <- confusionMatrix(
smile__svm_model_7J.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 16
## deliberate 14 16
##
## Accuracy : 0.5385
## 95% CI : (0.4103, 0.663)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.3553
##
## Kappa : 0.0758
##
## Mcnemar's Test P-Value : 0.8551
##
## Sensitivity : 0.5758
## Specificity : 0.5000
## Pos Pred Value : 0.5429
## Neg Pred Value : 0.5333
## Prevalence : 0.5077
## Detection Rate : 0.2923
## Detection Prevalence : 0.5385
## Balanced Accuracy : 0.5379
##
## 'Positive' Class : spontaneous
##
# 7K
set.seed(1973)
smile__svm_model_7K <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
offset_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_7K$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6302306 0.2607177 0.0555369 0.1093455
summary(smile__svm_model_7K$finalModel)
## Length Class Mode
## 1 ksvm S4
smile__svm_model_7K_pred <- predict(smile__svm_model_7K, tst_smile)
summary(smile__svm_model_7K_pred)
## spontaneous deliberate
## 64 78
smile__svm_model_7K_confM <- confusionMatrix(
smile__svm_model_7K_pred,
tst_smile$smile_type
)
smile__svm_model_7K_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 42 22
## deliberate 28 50
##
## Accuracy : 0.6479
## 95% CI : (0.5634, 0.7261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0004898
##
## Kappa : 0.2948
##
## Mcnemar's Test P-Value : 0.4795001
##
## Sensitivity : 0.6000
## Specificity : 0.6944
## Pos Pred Value : 0.6562
## Neg Pred Value : 0.6410
## Prevalence : 0.4930
## Detection Rate : 0.2958
## Detection Prevalence : 0.4507
## Balanced Accuracy : 0.6472
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_7K.1_pred <- predict(smile__svm_model_7K, tst_smile_boys)
summary(smile__svm_model_7K.1_pred)
## spontaneous deliberate
## 30 47
smile__svm_model_7K.1_confM <- confusionMatrix(
smile__svm_model_7K.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_7K.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 8
## deliberate 15 32
##
## Accuracy : 0.7013
## 95% CI : (0.5862, 0.8003)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.0008966
##
## Kappa : 0.3974
##
## Mcnemar's Test P-Value : 0.2109029
##
## Sensitivity : 0.5946
## Specificity : 0.8000
## Pos Pred Value : 0.7333
## Neg Pred Value : 0.6809
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.3896
## Balanced Accuracy : 0.6973
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_7K.2_pred <- predict(smile__svm_model_7K, tst_smile_girls)
summary(smile__svm_model_7K.2_pred)
## spontaneous deliberate
## 34 31
smile__svm_model_7K.2_confM <- confusionMatrix(
smile__svm_model_7K.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_7K.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 14
## deliberate 13 18
##
## Accuracy : 0.5846
## 95% CI : (0.4556, 0.7056)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.132
##
## Kappa : 0.1686
##
## Mcnemar's Test P-Value : 1.000
##
## Sensitivity : 0.6061
## Specificity : 0.5625
## Pos Pred Value : 0.5882
## Neg Pred Value : 0.5806
## Prevalence : 0.5077
## Detection Rate : 0.3077
## Detection Prevalence : 0.5231
## Balanced Accuracy : 0.5843
##
## 'Positive' Class : spontaneous
##
# 8 strongest features
# 8A
set.seed(1973)
smile__svm_model_8A <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean + lip_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8A$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6927028 0.3855624 0.08658439 0.1737373
smile__svm_model_8A_pred <- predict(smile__svm_model_8A, tst_smile)
summary(smile__svm_model_8A_pred)
## spontaneous deliberate
## 68 74
smile__svm_model_8A_confM <- confusionMatrix(
smile__svm_model_8A_pred,
tst_smile$smile_type
)
smile__svm_model_8A_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 24
## deliberate 26 48
##
## Accuracy : 0.6479
## 95% CI : (0.5634, 0.7261)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0004898
##
## Kappa : 0.2954
##
## Mcnemar's Test P-Value : 0.8875371
##
## Sensitivity : 0.6286
## Specificity : 0.6667
## Pos Pred Value : 0.6471
## Neg Pred Value : 0.6486
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.4789
## Balanced Accuracy : 0.6476
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8A.1_pred <- predict(smile__svm_model_8A, tst_smile_boys)
summary(smile__svm_model_8A.1_pred)
## spontaneous deliberate
## 27 50
smile__svm_model_8A.1_confM <- confusionMatrix(
smile__svm_model_8A.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8A.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 8
## deliberate 18 32
##
## Accuracy : 0.6623
## 95% CI : (0.5455, 0.7662)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.007868
##
## Kappa : 0.3167
##
## Mcnemar's Test P-Value : 0.077556
##
## Sensitivity : 0.5135
## Specificity : 0.8000
## Pos Pred Value : 0.7037
## Neg Pred Value : 0.6400
## Prevalence : 0.4805
## Detection Rate : 0.2468
## Detection Prevalence : 0.3506
## Balanced Accuracy : 0.6568
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8A.2_pred <- predict(smile__svm_model_8A, tst_smile_girls)
summary(smile__svm_model_8A.2_pred)
## spontaneous deliberate
## 41 24
smile__svm_model_8A.2_confM <- confusionMatrix(
smile__svm_model_8A.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8A.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 16
## deliberate 8 16
##
## Accuracy : 0.6308
## 95% CI : (0.502, 0.7472)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.03086
##
## Kappa : 0.2586
##
## Mcnemar's Test P-Value : 0.15304
##
## Sensitivity : 0.7576
## Specificity : 0.5000
## Pos Pred Value : 0.6098
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.3846
## Detection Prevalence : 0.6308
## Balanced Accuracy : 0.6288
##
## 'Positive' Class : spontaneous
##
# 8B
set.seed(1973)
smile__svm_model_8B <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + lip_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8B$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6633857 0.327155 0.06435706 0.12877
smile__svm_model_8B_pred <- predict(smile__svm_model_8B, tst_smile)
summary(smile__svm_model_8B_pred)
## spontaneous deliberate
## 74 68
smile__svm_model_8B_confM <- confusionMatrix(
smile__svm_model_8B_pred,
tst_smile$smile_type
)
smile__svm_model_8B_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 30
## deliberate 26 42
##
## Accuracy : 0.6056
## 95% CI : (0.5202, 0.6865)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.01151
##
## Kappa : 0.2117
##
## Mcnemar's Test P-Value : 0.68850
##
## Sensitivity : 0.6286
## Specificity : 0.5833
## Pos Pred Value : 0.5946
## Neg Pred Value : 0.6176
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.5211
## Balanced Accuracy : 0.6060
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8B.1_pred <- predict(smile__svm_model_8B, tst_smile_boys)
summary(smile__svm_model_8B.1_pred)
## spontaneous deliberate
## 34 43
smile__svm_model_8B.1_confM <- confusionMatrix(
smile__svm_model_8B.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8B.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 13
## deliberate 16 27
##
## Accuracy : 0.6234
## 95% CI : (0.5056, 0.7313)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.04297
##
## Kappa : 0.2433
##
## Mcnemar's Test P-Value : 0.71035
##
## Sensitivity : 0.5676
## Specificity : 0.6750
## Pos Pred Value : 0.6176
## Neg Pred Value : 0.6279
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.4416
## Balanced Accuracy : 0.6213
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8B.2_pred <- predict(smile__svm_model_8B, tst_smile_girls)
summary(smile__svm_model_8B.2_pred)
## spontaneous deliberate
## 40 25
smile__svm_model_8B.2_confM <- confusionMatrix(
smile__svm_model_8B.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8B.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 17
## deliberate 10 15
##
## Accuracy : 0.5846
## 95% CI : (0.4556, 0.7056)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.1320
##
## Kappa : 0.1663
##
## Mcnemar's Test P-Value : 0.2482
##
## Sensitivity : 0.6970
## Specificity : 0.4688
## Pos Pred Value : 0.5750
## Neg Pred Value : 0.6000
## Prevalence : 0.5077
## Detection Rate : 0.3538
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.5829
##
## 'Positive' Class : spontaneous
##
# 8C
set.seed(1973)
smile__svm_model_8C <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + apex_mean + lip_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8C$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5919285 0.1842753 0.09834628 0.1972986
smile__svm_model_8C_pred <- predict(smile__svm_model_8C, tst_smile)
summary(smile__svm_model_8C_pred)
## spontaneous deliberate
## 71 71
smile__svm_model_8C_confM <- confusionMatrix(
smile__svm_model_8C_pred,
tst_smile$smile_type
)
smile__svm_model_8C_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 41 30
## deliberate 29 42
##
## Accuracy : 0.5845
## 95% CI : (0.4989, 0.6665)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.03874
##
## Kappa : 0.169
##
## Mcnemar's Test P-Value : 1.00000
##
## Sensitivity : 0.5857
## Specificity : 0.5833
## Pos Pred Value : 0.5775
## Neg Pred Value : 0.5915
## Prevalence : 0.4930
## Detection Rate : 0.2887
## Detection Prevalence : 0.5000
## Balanced Accuracy : 0.5845
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8C.1_pred <- predict(smile__svm_model_8C, tst_smile_boys)
summary(smile__svm_model_8C.1_pred)
## spontaneous deliberate
## 34 43
smile__svm_model_8C.1_confM <- confusionMatrix(
smile__svm_model_8C.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8C.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 15
## deliberate 18 25
##
## Accuracy : 0.5714
## 95% CI : (0.4535, 0.6837)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.2126
##
## Kappa : 0.1389
##
## Mcnemar's Test P-Value : 0.7277
##
## Sensitivity : 0.5135
## Specificity : 0.6250
## Pos Pred Value : 0.5588
## Neg Pred Value : 0.5814
## Prevalence : 0.4805
## Detection Rate : 0.2468
## Detection Prevalence : 0.4416
## Balanced Accuracy : 0.5693
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8C.2_pred <- predict(smile__svm_model_8C, tst_smile_girls)
summary(smile__svm_model_8C.2_pred)
## spontaneous deliberate
## 37 28
smile__svm_model_8C.2_confM <- confusionMatrix(
smile__svm_model_8C.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8C.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 15
## deliberate 11 17
##
## Accuracy : 0.6
## 95% CI : (0.471, 0.7196)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.08588
##
## Kappa : 0.1983
##
## Mcnemar's Test P-Value : 0.55630
##
## Sensitivity : 0.6667
## Specificity : 0.5312
## Pos Pred Value : 0.5946
## Neg Pred Value : 0.6071
## Prevalence : 0.5077
## Detection Rate : 0.3385
## Detection Prevalence : 0.5692
## Balanced Accuracy : 0.5990
##
## 'Positive' Class : spontaneous
##
# 8D
set.seed(1973)
smile__svm_model_8D <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + offset_mean + lip_mean +
eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8D$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6632074 0.3276323 0.06838694 0.1362339
smile__svm_model_8D_pred <- predict(smile__svm_model_8D, tst_smile)
summary(smile__svm_model_8D_pred)
## spontaneous deliberate
## 78 64
smile__svm_model_8D_confM <- confusionMatrix(
smile__svm_model_8D_pred,
tst_smile$smile_type
)
smile__svm_model_8D_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 52 26
## deliberate 18 46
##
## Accuracy : 0.6901
## 95% CI : (0.6072, 0.765)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 7.345e-06
##
## Kappa : 0.3811
##
## Mcnemar's Test P-Value : 0.2913
##
## Sensitivity : 0.7429
## Specificity : 0.6389
## Pos Pred Value : 0.6667
## Neg Pred Value : 0.7188
## Prevalence : 0.4930
## Detection Rate : 0.3662
## Detection Prevalence : 0.5493
## Balanced Accuracy : 0.6909
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8D.1_pred <- predict(smile__svm_model_8D, tst_smile_boys)
summary(smile__svm_model_8D.1_pred)
## spontaneous deliberate
## 32 45
smile__svm_model_8D.1_confM <- confusionMatrix(
smile__svm_model_8D.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8D.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 9
## deliberate 14 31
##
## Accuracy : 0.7013
## 95% CI : (0.5862, 0.8003)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.0008966
##
## Kappa : 0.3986
##
## Mcnemar's Test P-Value : 0.4042485
##
## Sensitivity : 0.6216
## Specificity : 0.7750
## Pos Pred Value : 0.7188
## Neg Pred Value : 0.6889
## Prevalence : 0.4805
## Detection Rate : 0.2987
## Detection Prevalence : 0.4156
## Balanced Accuracy : 0.6983
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8D.2_pred <- predict(smile__svm_model_8D, tst_smile_girls)
summary(smile__svm_model_8D.2_pred)
## spontaneous deliberate
## 46 19
smile__svm_model_8D.2_confM <- confusionMatrix(
smile__svm_model_8D.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8D.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 29 17
## deliberate 4 15
##
## Accuracy : 0.6769
## 95% CI : (0.5495, 0.7877)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.004282
##
## Kappa : 0.3497
##
## Mcnemar's Test P-Value : 0.008829
##
## Sensitivity : 0.8788
## Specificity : 0.4688
## Pos Pred Value : 0.6304
## Neg Pred Value : 0.7895
## Prevalence : 0.5077
## Detection Rate : 0.4462
## Detection Prevalence : 0.7077
## Balanced Accuracy : 0.6738
##
## 'Positive' Class : spontaneous
##
# 8E
set.seed(1973)
smile__svm_model_8E <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean + lip_mean +
eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8E$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.7105336 0.4206647 0.1096118 0.2191265
smile__svm_model_8E_pred <- predict(smile__svm_model_8E, tst_smile)
summary(smile__svm_model_8E_pred)
## spontaneous deliberate
## 61 81
smile__svm_model_8E_confM <- confusionMatrix(
smile__svm_model_8E_pred,
tst_smile$smile_type
)
smile__svm_model_8E_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 47 14
## deliberate 23 58
##
## Accuracy : 0.7394
## 95% CI : (0.6592, 0.8094)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 1.288e-08
##
## Kappa : 0.4778
##
## Mcnemar's Test P-Value : 0.1884
##
## Sensitivity : 0.6714
## Specificity : 0.8056
## Pos Pred Value : 0.7705
## Neg Pred Value : 0.7160
## Prevalence : 0.4930
## Detection Rate : 0.3310
## Detection Prevalence : 0.4296
## Balanced Accuracy : 0.7385
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8E.1_pred <- predict(smile__svm_model_8E, tst_smile_boys)
summary(smile__svm_model_8E.1_pred)
## spontaneous deliberate
## 25 52
smile__svm_model_8E.1_confM <- confusionMatrix(
smile__svm_model_8E.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8E.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 4
## deliberate 16 36
##
## Accuracy : 0.7403
## 95% CI : (0.6277, 0.8336)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 6.073e-05
##
## Kappa : 0.4733
##
## Mcnemar's Test P-Value : 0.01391
##
## Sensitivity : 0.5676
## Specificity : 0.9000
## Pos Pred Value : 0.8400
## Neg Pred Value : 0.6923
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.3247
## Balanced Accuracy : 0.7338
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8E.2_pred <- predict(smile__svm_model_8E, tst_smile_girls)
summary(smile__svm_model_8E.2_pred)
## spontaneous deliberate
## 36 29
smile__svm_model_8E.2_confM <- confusionMatrix(
smile__svm_model_8E.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8E.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 10
## deliberate 7 22
##
## Accuracy : 0.7385
## 95% CI : (0.6146, 0.8397)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.0001234
##
## Kappa : 0.4761
##
## Mcnemar's Test P-Value : 0.6276258
##
## Sensitivity : 0.7879
## Specificity : 0.6875
## Pos Pred Value : 0.7222
## Neg Pred Value : 0.7586
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.5538
## Balanced Accuracy : 0.7377
##
## 'Positive' Class : spontaneous
##
# 8F
set.seed(1973)
smile__svm_model_8F <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8F$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6602607 0.3204802 0.09509904 0.1915229
smile__svm_model_8F_pred <- predict(smile__svm_model_8F, tst_smile)
summary(smile__svm_model_8F_pred)
## spontaneous deliberate
## 64 78
smile__svm_model_8F_confM <- confusionMatrix(
smile__svm_model_8F_pred,
tst_smile$smile_type
)
smile__svm_model_8F_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 43 21
## deliberate 27 51
##
## Accuracy : 0.662
## 95% CI : (0.5779, 0.7391)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.0001362
##
## Kappa : 0.323
##
## Mcnemar's Test P-Value : 0.4704864
##
## Sensitivity : 0.6143
## Specificity : 0.7083
## Pos Pred Value : 0.6719
## Neg Pred Value : 0.6538
## Prevalence : 0.4930
## Detection Rate : 0.3028
## Detection Prevalence : 0.4507
## Balanced Accuracy : 0.6613
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8F.1_pred <- predict(smile__svm_model_8F, tst_smile_boys)
summary(smile__svm_model_8F.1_pred)
## spontaneous deliberate
## 26 51
smile__svm_model_8F.1_confM <- confusionMatrix(
smile__svm_model_8F.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8F.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 19 7
## deliberate 18 33
##
## Accuracy : 0.6753
## 95% CI : (0.559, 0.7777)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.00403
##
## Kappa : 0.3423
##
## Mcnemar's Test P-Value : 0.04550
##
## Sensitivity : 0.5135
## Specificity : 0.8250
## Pos Pred Value : 0.7308
## Neg Pred Value : 0.6471
## Prevalence : 0.4805
## Detection Rate : 0.2468
## Detection Prevalence : 0.3377
## Balanced Accuracy : 0.6693
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8F.2_pred <- predict(smile__svm_model_8F, tst_smile_girls)
summary(smile__svm_model_8F.2_pred)
## spontaneous deliberate
## 38 27
smile__svm_model_8F.2_confM <- confusionMatrix(
smile__svm_model_8F.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8F.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 14
## deliberate 9 18
##
## Accuracy : 0.6462
## 95% CI : (0.5177, 0.7608)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.01698
##
## Kappa : 0.2905
##
## Mcnemar's Test P-Value : 0.40425
##
## Sensitivity : 0.7273
## Specificity : 0.5625
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.5846
## Balanced Accuracy : 0.6449
##
## 'Positive' Class : spontaneous
##
# 8G
set.seed(1973)
smile__svm_model_8G <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8G$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5408422 0.08175117 0.04353818 0.08936622
smile__svm_model_8G_pred <- predict(smile__svm_model_8G, tst_smile)
summary(smile__svm_model_8G_pred)
## spontaneous deliberate
## 73 69
smile__svm_model_8G_confM <- confusionMatrix(
smile__svm_model_8G_pred,
tst_smile$smile_type
)
smile__svm_model_8G_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 42 31
## deliberate 28 41
##
## Accuracy : 0.5845
## 95% CI : (0.4989, 0.6665)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.03874
##
## Kappa : 0.1693
##
## Mcnemar's Test P-Value : 0.79457
##
## Sensitivity : 0.6000
## Specificity : 0.5694
## Pos Pred Value : 0.5753
## Neg Pred Value : 0.5942
## Prevalence : 0.4930
## Detection Rate : 0.2958
## Detection Prevalence : 0.5141
## Balanced Accuracy : 0.5847
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8G.1_pred <- predict(smile__svm_model_8G, tst_smile_boys)
summary(smile__svm_model_8G.1_pred)
## spontaneous deliberate
## 33 44
smile__svm_model_8G.1_confM <- confusionMatrix(
smile__svm_model_8G.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8G.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 17 16
## deliberate 20 24
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.0597
##
## Mcnemar's Test P-Value : 0.6171
##
## Sensitivity : 0.4595
## Specificity : 0.6000
## Pos Pred Value : 0.5152
## Neg Pred Value : 0.5455
## Prevalence : 0.4805
## Detection Rate : 0.2208
## Detection Prevalence : 0.4286
## Balanced Accuracy : 0.5297
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8G.2_pred <- predict(smile__svm_model_8G, tst_smile_girls)
summary(smile__svm_model_8G.2_pred)
## spontaneous deliberate
## 40 25
smile__svm_model_8G.2_confM <- confusionMatrix(
smile__svm_model_8G.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8G.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 15
## deliberate 8 17
##
## Accuracy : 0.6462
## 95% CI : (0.5177, 0.7608)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.01698
##
## Kappa : 0.2898
##
## Mcnemar's Test P-Value : 0.21090
##
## Sensitivity : 0.7576
## Specificity : 0.5312
## Pos Pred Value : 0.6250
## Neg Pred Value : 0.6800
## Prevalence : 0.5077
## Detection Rate : 0.3846
## Detection Prevalence : 0.6154
## Balanced Accuracy : 0.6444
##
## 'Positive' Class : spontaneous
##
# 8H
set.seed(1973)
smile__svm_model_8H <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + apex_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8H$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5591188 0.118238 0.06334477 0.1273335
smile__svm_model_8H_pred <- predict(smile__svm_model_8H, tst_smile)
summary(smile__svm_model_8H_pred)
## spontaneous deliberate
## 66 76
smile__svm_model_8H_confM <- confusionMatrix(
smile__svm_model_8H_pred,
tst_smile$smile_type
)
smile__svm_model_8H_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 36 30
## deliberate 34 42
##
## Accuracy : 0.5493
## 95% CI : (0.4636, 0.6328)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.1780
##
## Kappa : 0.0977
##
## Mcnemar's Test P-Value : 0.7077
##
## Sensitivity : 0.5143
## Specificity : 0.5833
## Pos Pred Value : 0.5455
## Neg Pred Value : 0.5526
## Prevalence : 0.4930
## Detection Rate : 0.2535
## Detection Prevalence : 0.4648
## Balanced Accuracy : 0.5488
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8H.1_pred <- predict(smile__svm_model_8H, tst_smile_boys)
summary(smile__svm_model_8H.1_pred)
## spontaneous deliberate
## 27 50
smile__svm_model_8H.1_confM <- confusionMatrix(
smile__svm_model_8H.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8H.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 16 11
## deliberate 21 29
##
## Accuracy : 0.5844
## 95% CI : (0.4664, 0.6957)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.1523
##
## Kappa : 0.159
##
## Mcnemar's Test P-Value : 0.1116
##
## Sensitivity : 0.4324
## Specificity : 0.7250
## Pos Pred Value : 0.5926
## Neg Pred Value : 0.5800
## Prevalence : 0.4805
## Detection Rate : 0.2078
## Detection Prevalence : 0.3506
## Balanced Accuracy : 0.5787
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8H.2_pred <- predict(smile__svm_model_8H, tst_smile_girls)
summary(smile__svm_model_8H.2_pred)
## spontaneous deliberate
## 39 26
smile__svm_model_8H.2_confM <- confusionMatrix(
smile__svm_model_8H.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8H.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 20 19
## deliberate 13 13
##
## Accuracy : 0.5077
## 95% CI : (0.3807, 0.634)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.5495
##
## Kappa : 0.0123
##
## Mcnemar's Test P-Value : 0.3768
##
## Sensitivity : 0.6061
## Specificity : 0.4062
## Pos Pred Value : 0.5128
## Neg Pred Value : 0.5000
## Prevalence : 0.5077
## Detection Rate : 0.3077
## Detection Prevalence : 0.6000
## Balanced Accuracy : 0.5062
##
## 'Positive' Class : spontaneous
##
# 8I
set.seed(1973)
smile__svm_model_8I <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + offset_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8I$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5761308 0.152213 0.07350272 0.1491875
smile__svm_model_8I_pred <- predict(smile__svm_model_8I, tst_smile)
summary(smile__svm_model_8I_pred)
## spontaneous deliberate
## 67 75
smile__svm_model_8I_confM <- confusionMatrix(
smile__svm_model_8I_pred,
tst_smile$smile_type
)
smile__svm_model_8I_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 39 28
## deliberate 31 44
##
## Accuracy : 0.5845
## 95% CI : (0.4989, 0.6665)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 0.03874
##
## Kappa : 0.1684
##
## Mcnemar's Test P-Value : 0.79457
##
## Sensitivity : 0.5571
## Specificity : 0.6111
## Pos Pred Value : 0.5821
## Neg Pred Value : 0.5867
## Prevalence : 0.4930
## Detection Rate : 0.2746
## Detection Prevalence : 0.4718
## Balanced Accuracy : 0.5841
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8I.1_pred <- predict(smile__svm_model_8I, tst_smile_boys)
summary(smile__svm_model_8I.1_pred)
## spontaneous deliberate
## 29 48
smile__svm_model_8I.1_confM <- confusionMatrix(
smile__svm_model_8I.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8I.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 15 14
## deliberate 22 26
##
## Accuracy : 0.5325
## 95% CI : (0.4152, 0.6471)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.4552
##
## Kappa : 0.0559
##
## Mcnemar's Test P-Value : 0.2433
##
## Sensitivity : 0.4054
## Specificity : 0.6500
## Pos Pred Value : 0.5172
## Neg Pred Value : 0.5417
## Prevalence : 0.4805
## Detection Rate : 0.1948
## Detection Prevalence : 0.3766
## Balanced Accuracy : 0.5277
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8I.2_pred <- predict(smile__svm_model_8I, tst_smile_girls)
summary(smile__svm_model_8I.2_pred)
## spontaneous deliberate
## 38 27
smile__svm_model_8I.2_confM <- confusionMatrix(
smile__svm_model_8I.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8I.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 24 14
## deliberate 9 18
##
## Accuracy : 0.6462
## 95% CI : (0.5177, 0.7608)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.01698
##
## Kappa : 0.2905
##
## Mcnemar's Test P-Value : 0.40425
##
## Sensitivity : 0.7273
## Specificity : 0.5625
## Pos Pred Value : 0.6316
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.3692
## Detection Prevalence : 0.5846
## Balanced Accuracy : 0.6449
##
## 'Positive' Class : spontaneous
##
# 8J
set.seed(1973)
smile__svm_model_8J <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean + onset_mean + apex_mean +
offset_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8J$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6926081 0.3854562 0.0926346 0.1859359
smile__svm_model_8J_pred <- predict(smile__svm_model_8J, tst_smile)
summary(smile__svm_model_8J_pred)
## spontaneous deliberate
## 69 73
smile__svm_model_8J_confM <- confusionMatrix(
smile__svm_model_8J_pred,
tst_smile$smile_type
)
smile__svm_model_8J_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 47 22
## deliberate 23 50
##
## Accuracy : 0.6831
## 95% CI : (0.5998, 0.7586)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 1.596e-05
##
## Kappa : 0.3659
##
## Mcnemar's Test P-Value : 1
##
## Sensitivity : 0.6714
## Specificity : 0.6944
## Pos Pred Value : 0.6812
## Neg Pred Value : 0.6849
## Prevalence : 0.4930
## Detection Rate : 0.3310
## Detection Prevalence : 0.4859
## Balanced Accuracy : 0.6829
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8J.1_pred <- predict(smile__svm_model_8J, tst_smile_boys)
summary(smile__svm_model_8J.1_pred)
## spontaneous deliberate
## 30 47
smile__svm_model_8J.1_confM <- confusionMatrix(
smile__svm_model_8J.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8J.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 8
## deliberate 15 32
##
## Accuracy : 0.7013
## 95% CI : (0.5862, 0.8003)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 0.0008966
##
## Kappa : 0.3974
##
## Mcnemar's Test P-Value : 0.2109029
##
## Sensitivity : 0.5946
## Specificity : 0.8000
## Pos Pred Value : 0.7333
## Neg Pred Value : 0.6809
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.3896
## Balanced Accuracy : 0.6973
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8J.2_pred <- predict(smile__svm_model_8J, tst_smile_girls)
summary(smile__svm_model_8J.2_pred)
## spontaneous deliberate
## 39 26
smile__svm_model_8J.2_confM <- confusionMatrix(
smile__svm_model_8J.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8J.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 25 14
## deliberate 8 18
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.321
##
## Mcnemar's Test P-Value : 0.286422
##
## Sensitivity : 0.7576
## Specificity : 0.5625
## Pos Pred Value : 0.6410
## Neg Pred Value : 0.6923
## Prevalence : 0.5077
## Detection Rate : 0.3846
## Detection Prevalence : 0.6000
## Balanced Accuracy : 0.6600
##
## 'Positive' Class : spontaneous
##
# 8K
set.seed(1973)
smile__svm_model_8K <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean + lip_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8K$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.7196301 0.4390769 0.1031942 0.205939
smile__svm_model_8K_pred <- predict(smile__svm_model_8K, tst_smile)
summary(smile__svm_model_8K_pred)
## spontaneous deliberate
## 63 79
smile__svm_model_8K_confM <- confusionMatrix(
smile__svm_model_8K_pred,
tst_smile$smile_type
)
smile__svm_model_8K_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 48 15
## deliberate 22 57
##
## Accuracy : 0.7394
## 95% CI : (0.6592, 0.8094)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 1.288e-08
##
## Kappa : 0.478
##
## Mcnemar's Test P-Value : 0.3239
##
## Sensitivity : 0.6857
## Specificity : 0.7917
## Pos Pred Value : 0.7619
## Neg Pred Value : 0.7215
## Prevalence : 0.4930
## Detection Rate : 0.3380
## Detection Prevalence : 0.4437
## Balanced Accuracy : 0.7387
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8K.1_pred <- predict(smile__svm_model_8K, tst_smile_boys)
summary(smile__svm_model_8K.1_pred)
## spontaneous deliberate
## 26 51
smile__svm_model_8K.1_confM <- confusionMatrix(
smile__svm_model_8K.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8K.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 22 4
## deliberate 15 36
##
## Accuracy : 0.7532
## 95% CI : (0.6418, 0.8444)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 2.185e-05
##
## Kappa : 0.5002
##
## Mcnemar's Test P-Value : 0.02178
##
## Sensitivity : 0.5946
## Specificity : 0.9000
## Pos Pred Value : 0.8462
## Neg Pred Value : 0.7059
## Prevalence : 0.4805
## Detection Rate : 0.2857
## Detection Prevalence : 0.3377
## Balanced Accuracy : 0.7473
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8K.2_pred <- predict(smile__svm_model_8K, tst_smile_girls)
summary(smile__svm_model_8K.2_pred)
## spontaneous deliberate
## 37 28
smile__svm_model_8K.2_confM <- confusionMatrix(
smile__svm_model_8K.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8K.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 26 11
## deliberate 7 21
##
## Accuracy : 0.7231
## 95% CI : (0.5981, 0.8269)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.0003328
##
## Kappa : 0.445
##
## Mcnemar's Test P-Value : 0.4795001
##
## Sensitivity : 0.7879
## Specificity : 0.6562
## Pos Pred Value : 0.7027
## Neg Pred Value : 0.7500
## Prevalence : 0.5077
## Detection Rate : 0.4000
## Detection Prevalence : 0.5692
## Balanced Accuracy : 0.7221
##
## 'Positive' Class : spontaneous
##
# 8L
set.seed(1973)
smile__svm_model_8L <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
onset_mean + apex_mean + offset_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_8L$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6897616 0.3788924 0.08931016 0.1785103
smile__svm_model_8L_pred <- predict(smile__svm_model_8L, tst_smile)
summary(smile__svm_model_8L_pred)
## spontaneous deliberate
## 60 82
smile__svm_model_8L_confM <- confusionMatrix(
smile__svm_model_8L_pred,
tst_smile$smile_type
)
smile__svm_model_8L_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 44 16
## deliberate 26 56
##
## Accuracy : 0.7042
## 95% CI : (0.6219, 0.7778)
## No Information Rate : 0.507
## P-Value [Acc > NIR] : 1.414e-06
##
## Kappa : 0.4072
##
## Mcnemar's Test P-Value : 0.1649
##
## Sensitivity : 0.6286
## Specificity : 0.7778
## Pos Pred Value : 0.7333
## Neg Pred Value : 0.6829
## Prevalence : 0.4930
## Detection Rate : 0.3099
## Detection Prevalence : 0.4225
## Balanced Accuracy : 0.7032
##
## 'Positive' Class : spontaneous
##
# predicting boys, girls
set.seed(1973)
smile__svm_model_8L.1_pred <- predict(smile__svm_model_8L, tst_smile_boys)
summary(smile__svm_model_8L.1_pred)
## spontaneous deliberate
## 25 52
smile__svm_model_8L.1_confM <- confusionMatrix(
smile__svm_model_8L.1_pred,
tst_smile_boys$smile_type
)
smile__svm_model_8L.1_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 21 4
## deliberate 16 36
##
## Accuracy : 0.7403
## 95% CI : (0.6277, 0.8336)
## No Information Rate : 0.5195
## P-Value [Acc > NIR] : 6.073e-05
##
## Kappa : 0.4733
##
## Mcnemar's Test P-Value : 0.01391
##
## Sensitivity : 0.5676
## Specificity : 0.9000
## Pos Pred Value : 0.8400
## Neg Pred Value : 0.6923
## Prevalence : 0.4805
## Detection Rate : 0.2727
## Detection Prevalence : 0.3247
## Balanced Accuracy : 0.7338
##
## 'Positive' Class : spontaneous
##
set.seed(1973)
smile__svm_model_8L.2_pred <- predict(smile__svm_model_8L, tst_smile_girls)
summary(smile__svm_model_8L.2_pred)
## spontaneous deliberate
## 35 30
smile__svm_model_8L.2_confM <- confusionMatrix(
smile__svm_model_8L.2_pred,
tst_smile_girls$smile_type
)
smile__svm_model_8L.2_confM
## Confusion Matrix and Statistics
##
## Reference
## Prediction spontaneous deliberate
## spontaneous 23 12
## deliberate 10 20
##
## Accuracy : 0.6615
## 95% CI : (0.5335, 0.7743)
## No Information Rate : 0.5077
## P-Value [Acc > NIR] : 0.008794
##
## Kappa : 0.3223
##
## Mcnemar's Test P-Value : 0.831170
##
## Sensitivity : 0.6970
## Specificity : 0.6250
## Pos Pred Value : 0.6571
## Neg Pred Value : 0.6667
## Prevalence : 0.5077
## Detection Rate : 0.3538
## Detection Prevalence : 0.5385
## Balanced Accuracy : 0.6610
##
## 'Positive' Class : spontaneous
##
# test two models without temporal features
# they do not improve the accuracy scores
set.seed(1973)
smile__svm_model_9 <- train(smile_type ~ AU01_r_mean + AU09_r_mean +
AU10_r_mean + AU25_r_mean + AU45_r_mean +
lip_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_9$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.6362021 0.2726033 0.0691034 0.1362343
set.seed(1973)
smile__svm_model_10 <- train(smile_type ~ AU06_r_mean + AU12_r_mean +
AU45_r_mean +
lip_mean + eye_mean,
method = "svmLinear", data = trn_smile,
trControl = trainControl(method = "cv", number = 10)
)
smile__svm_model_10$results
## C Accuracy Kappa AccuracySD KappaSD
## 1 1 0.5677585 0.1365759 0.05687368 0.1156556
In the exploration phase of the thesis the OpenFacer package was tested. This package published on GitHub, needed installation via the devtools package. The installation code is shown below.
# install_github("davidecannatanuig/openFaceR")
library(openfacer)
# citation("openfacer")
# library(devtools)
# citation("devtools")
Openfacer is a tool developed for social sientists. This tool is used to set up feature creation of this study and as it resources were to limited, the results on the descriptive statistics were compared. They show the same outcome, and the choice was made to move on with database which was build by myself, because of more flexibility in the analysis.
# loading packages
library(openfacer)
library(readr)
library(tidyverse)
# read the CSV files at once into RStudio
UvA_face <- read_face_csvs("Data_Openface/CSV/")
# create features using a pipeline
UvA_final_face <- UvA_face %>%
select_faces(
starts_with("gaze_"), starts_with("pose_"), starts_with("AU06"),
starts_with("AU12"), x_36, x_37, x_38, x_39, x_42, x_43, x_44, x_45,
x_48, x_54, y_36, y_37, y_38, y_39, y_42, y_43, y_44, y_45, y_48,
y_54
) %>%
mutate_faces(AU06_12_c = ifelse(AU06_c == 1 & AU12_c == 1, 1, 0)) %>%
mutate_faces(lip = sqrt((x_48 - x_54)^2 + (y_48 - y_54)^2)) %>%
mutate_faces(eye_x_m_l = (x_36 + x_39) / 2) %>%
mutate_faces(eye_y_m_l = (y_36 + y_39) / 2) %>%
mutate_faces(eye_x_m_r = (x_42 + x_45) / 2) %>%
mutate_faces(eye_y_m_r = (y_42 + y_45) / 2) %>%
mutate_faces(eye_x_u_l = (x_37 + x_38) / 2) %>%
mutate_faces(eye_y_u_l = (y_37 + y_38) / 2) %>%
mutate_faces(eye_x_u_r = (x_43 + x_44) / 2) %>%
mutate_faces(eye_y_u_r = (y_43 + y_44) / 2) %>%
mutate_faces(eye_l = sqrt((eye_x_m_l - eye_x_u_l)^2 +
(eye_y_m_l - eye_y_u_l)^2)) %>%
mutate_faces(eye_r = sqrt((eye_x_m_r - eye_x_u_r)^2 +
(eye_y_m_r - eye_y_u_r)^2)) %>%
mutate_faces(eye = (eye_l + eye_r) / 2) %>%
select_faces(
starts_with("gaze_"), starts_with("pose_"), starts_with("AU06"),
starts_with("AU12"), lip, eye_l, eye_r, eye
) %>%
tidy_face()
# save the data frame
write_csv(UvA_final_face, "UvA_final_face")
# create a check file to the own created summary file.
# compare two files have the same content using all()
UvA_face_check <- read.csv("UvA_final_face")
# all()
To support the thesis a figure is created displaying all the 2D landmark features. This is done based on subject number 20, as this subject is the only in database which data can be used in the report with permission of the data holder. The picture can be found in the thesis file.
library(ggplot2)
library(dplyr)
library(reshape2)
# citation("reshape2")
figure_1 <- read.csv("masterfile_connected")
figure_2 <- figure_1 %>%
select(starts_with("x_"))
figure_2 <- figure_2[50, ]
fig_3 <- dcast(melt(as.matrix(figure_2)),
Var2 ~ paste0("x", Var1),
value.var = "value"
)
figure_4 <- figure_1 %>%
select(starts_with("y_"))
figure_4 <- figure_4[50, ]
fig_5 <- dcast(melt(as.matrix(figure_4)),
Var2 ~ paste0("y", Var1),
value.var = "value"
)
fig_6 <- cbind(fig_5, fig_3)
fig_6 <- fig_6[1:68, ]
fig_6$Var2 <- gsub("y_", "", fig_6$Var2)
par(mfrow = c(1, 1))
dev.new(width = 5, height = 8)
plot(fig_6$x50, fig_6$y50,
xlim = c(950, 400), ylim = c(850, 300),
col = "blue",
xlab = "landmark x points", ylab = "landmark y points"
)
text(fig_6$x50, fig_6$y50, labels = fig_6$Var2, cex = .6, pos = 4)